---
title: 'Process Mining and Business Automation Full Course | Business Automation for Beginners | Simplilearn'
source: 'https://youtube.com/watch?v=B8q67cKoKY0'
video_id: 'B8q67cKoKY0'
date: 2026-08-08
duration_sec: 9351
---

# Process Mining and Business Automation Full Course | Business Automation for Beginners | Simplilearn

> Source: [Process Mining and Business Automation Full Course | Business Automation for Beginners | Simplilearn](https://youtube.com/watch?v=B8q67cKoKY0)

## Summary

This course provides a comprehensive introduction to business automation, process mining, and RPA, using real-world examples to illustrate concepts. It covers the role of a business analyst, creating BRDs, and practical applications like building a custom GPT for invoice analysis.

### Key Points

- **Automation's True Purpose** [00:02] — Automation is not about replacing people with tools but about identifying repeated tasks, improving workflows, and helping teams focus on work that requires human thinking.
- **Customer Feedback Example** [00:16] — A business receiving customer feedback from emails, forms, reviews, chat support, and social media can save hours by automating the classification, pattern finding, and report preparation.
- **Course Roadmap** [00:56] — The course covers basics of business automation, process standardization, process mining, RPA, workflows, business rules, and applications in customer feedback analysis, invoice processing, sales reporting, stock management, and decision making.
- **Business Analyst's Role** [01:25] — A business analyst understands the business problem, identifies stakeholders, studies the current process, defines requirements, checks risks, validates assumptions, and ensures the solution is useful.
- **BRD Components** [01:52] — A Business Requirement Document includes executive summary, project objectives, project scope, business requirements, stakeholder constraints, cost, and expected outcomes.
- **AI as a Support Tool** [02:07] — Tools like ChatGPT can create faster starting points for BRDs, but AI does not replace business understanding; the real value comes from validating the solution and ensuring practicality.
- **Best Fit for Automation** [04:10] — The best fit for automation is a repetitive rule-based task, not creative tasks, one-time discussions, or tasks with unclear requirements.
- **Interactive Session Start** [05:58] — The session is interactive, with participants from various sectors like banking and services, discussing their familiarity with terms like business automation, RPA, and process mining.
- **Problem Statement: Customer Feedback** [07:58] — A BA at an e-commerce company spends hours compiling customer feedback; the task is to think of a solution in layman terms.
- **Solution Ideas from Participants** [11:10] — Participants suggest quantifying and classifying data, setting parameters, using filters, identifying duplicates, and assigning weightage to keywords for trend analysis.
- **Automation Backfire Example** [16:38] — An Indigo customer posted a sarcastic tweet about a horrible experience; the bot interpreted positive words like 'amazing' and responded with encouragement, showing automation's limitations.
- **Snighel's Fruit Shop Example** [19:31] — A simple example of a fruit shop owner who manually copies invoice data to Excel; this is a standardized, rule-based task that can be automated.
- **Automation Benefits** [23:05] — Automation saves time and effort, provides accurate data, helps in understanding sales trends, and allows the owner to focus on business expansion.
- **Standardization as First Step** [26:40] — Standardization is the first step towards automation; if a process is straightforward and rule-based, it can be automated.
- **Convincing the Client** [28:51] — To convince a skeptical client, quantify benefits, provide a demo or prototype, and test in a small environment before scaling.
- **Documentation and BRD** [34:02] — Everything must be documented in a BRD to limit scope and ensure delivery; it prevents scope creep and sets clear expectations.
- **Decathlon Self-Checkout Example** [37:11] — Decathlon analyzed the manual checkout process (process mining) to create a self-checkout system using RFID scanners, illustrating how process mining works.
- **Validation with POC** [41:27] — In a real-world example, Accenture and Novartis share the cost of a proof of concept (POC) to validate an idea before scaling, reducing risk for the client.
- **BRD and Risk Mitigation** [44:41] — BRDs include risk identification and mitigation steps, ensuring business needs are clearly defined and communicated.
- **Creating a BRD with ChatGPT** [47:25] — The instructor demonstrates creating a BRD using ChatGPT by providing a problem statement and components, showing how AI can speed up the process.
- **BA's Role Beyond BRD** [53:23] — The BA's role is to put the correct prompt, take feedback, divide the scope into phases, and communicate delivery timelines; clients don't read the entire BRD.
- **BRD vs PRD** [56:46] — A PRD (Product Requirements Document) goes into nitty-gritty details like data sources and report components, while a BRD is high-level; both are used by BAs and PMs.
- **Creating an RPA with GPT** [01:00:57] — The instructor creates a custom GPT that extracts date and amount from invoice PDFs and creates a date-wise sales table, demonstrating a simple RPA.
- **Capstone Projects Overview** [01:10:44] — The course includes three capstone projects: credit card fraud analytics, patient readmission analytics, and retail analytics, each requiring BRD, SQL, Excel, and PowerBI tasks.
- **Credit Card Fraud Project** [01:11:15] — The project involves analyzing 42,000 transactions to identify fraud patterns using BRD, SQL, Excel, and PowerBI, with a focus on data-driven decisions.
- **Patient Readmission Project** [01:38:48] — The project aims to identify patient readmission risks using a dataset of 101,000 records, with tasks in BRD, SQL, Excel, and PowerBI for healthcare resource optimization.
- **Retail Analytics Project** [01:48:31] — The project focuses on predicting customer responses to marketing campaigns using a dataset of 2,200 customers, with tasks in BRD, SQL, Excel, and PowerBI for campaign optimization.

### Conclusion

Business automation starts with understanding and standardizing processes, then using tools like RPA and AI to streamline tasks. The role of a business analyst is crucial in documenting requirements, validating solutions, and ensuring practical delivery.

## Transcript

course by simply learn. Today every company wants to save time, reduce manual work, improve accuracy and make faster business decisions. But here's the important part. Automation is just not about replacing people with tools.
It's about identifying the repeated task, improving workflows, and helping teams focus on work that actually needs human thinking. Think about a simple example. a business that receives customer feedback every day from emails,
forms, reviews, chat support, and social media. And if a person has to read every comment manually, classify the complaint and find patterns and prepare report, it
can take hours and hours to do that. But with automation, this process can become faster, cleaner, and much easier to manage. Now, that's exactly what we are going to understand in this course. We will begin by the basics of business
automation and understand why companies automate repetitive and rule-based task. Then we are going to learn how process standardization becomes the first step towards automation. And after that we will explore important concepts like
process mining, RPA, workflows, business rules and how automation helps in areas like customer feedback analysis. Also in areas like invoice processing, sales reporting, stock management and business decision making. We are also going to
understand how a business analyst plays important role in automation projects. A business analyst does not just say let's automate this. They first understand the business problem, identify stakeholders, study the current process, define
requirements, check risk, validate assumptions and make sure the solution is useful for the business. Then we will move into practical areas like creating BRD or business requirement document where you're going to understand what
goes inside BD including executive summary, project objectives, project scope, business requirement, stakeholder constraints, cost and expected outcomes. We are going to see how tools like Chad Gibbit and other AI platforms can help
us create faster starting points. But remember guys, AI does not replace business understanding. It just supports it. The real value still comes us from the client's problem, validating the solution and making sure the final
output is very practical. So guys, if you want to know more about it, then watch this video till the end. Now before we begin our session, let's take a quick look at the program that can help you build skills in AI, machine
learning, and intelligent automation. If you're interested in building a strong career in AI and automation, the professional certificate program in intelligent automation by simply learn is a great opportunity for you. This is
11 month live and interactive program and it is powered by simply learn in collaboration with IHFC technology innovation hub of IIT Delhi. It is designed to help professionals build practical skills in AI, machine
learning, genative AI, deep learning and intelligent automation. You will learn important topics like Python, applied data science, machine learning, deep learning, genai literacy and many more amazing things. You will gain hands-on
experience with 20 plus AI tools and libraries including Python, Langchain, Chat, GPT, OpenAI, Hugging Phase, Gradio, TensorFlow, Scikitlearn and many more things. Apart from this, this program also includes 12 plus industry
aligning projects, 165 plus guided exercises, Microsoft course completion badges and trophies and also master classes by IHFC TIH of IIT Delhi faculty. And also you're going to get a 2-day campus immersion at IIT Delhi.
You're also going to get a job assist from simply learn to help you strengthen your professional profile and get noticed by top hiring companies. So if generative AI, machine learning and intelligent automation, check out the
course link in the description box and in the pin comments. Now before you begin this course, let's start with a quick question. And the question is which task is usually the best fit for automation? And your options are
creative task with no clear pattern, a repetitive rule-based task, one-time discussion, or a task with unclear requirements. Please mention your answers in the comment section below. learners let's say AJ Abini Aira Alisha
because you know when we begin with the session we discuss together it's always good to look at each other and then talk all right so today's session it is not at all a PPT based it's not at all related anything to theory we'll only be
discussing live examples from our experience as working professionals all right therefore it makes a lot of sense to look at each other and then talk all right so we have got seller partner from Chennai, Hyderabad, Bank of America. All
right. Kagi, you shifted from an from the services sector to the client side. Maybe from Infosys or Accenture or you joined Bank of America directly.
right, I see learners. We've got a good diversified pool. Okay, directly. We've got a good diversified pool of learners and we have joining from all over the Thank you Shika for switching on your video. I see lots of people joining from
south. [gasps] There are some people who are from north as well. Great. Great. All right. Sure. Learners, you know, I have a very simple quick question for you all. Right. If you could just share uh see my screen. I'm switching on my
whiteboard and a very small deck over here.
>> Yes. >> Has anyone heard of any of these three terms? Let's say you know if I ask you as a layman, people from bank also would also be well aware these days. So anyone who has heard okay Supa said no s I've
heard of these terms and maybe you would like to share one or two lines about what you know about that's it. >> Business automation is basically the way the operation. >> Uhhuh. Uhhuh.
Okay. Siddesh we've got someone joining from all state also. Okay. Pune. Great. Thank you. Sesh. Anyone? Santo is there as partner. >> So on Yeah. So I have worked on a tool called as blue prism. So which is
>> RPA tool. >> I I was actually just involved in making it. >> Okay. >> you have also worked with automation
anywhere by any chance? >> Uh not automation anywhere but uh otherwise like business process uh like BPSM tools like Appen I have worked with. So I was a developer few years back and right now uh in my current
organizations like we are trying to build an in-house tool uh which is which works very similar to appen like where you can like drag and drop components and create workflows. >> Got it. Got it. Something similar to
Zapier. Okay. Thank you Nikl. Uh next next take. you know why don't you pitch in if you've heard of even process mining is fine maybe someone from bank we got two three people from the banking sector
right they would be knowing much better these days about RPA or business automation or process mining anyone would like to pitch in
with very simple problem statement over here all right I'll give you a minute read this and then I'll get back. I'll give you a minute to read. All right. Once you have read, why don't you just drop a yes in
the chat box. All right. So, just take a minute. It says you're a BA. You're working at an e-commerce. Uh, you know, your team spends hours compiling your team spends hours compiling customer feedback.
Great. Now learners what we'll do is I'll keep the problem statement open for you all to see. All right. And I'll give you another 1 minute. What you need to do is very simple layman terms. Maybe you know you do it on a daily basis in
your sector. For example, as an R as RM, you would be taking lots and lots of let's say the app isn't working, this feature is not working. Sometimes people would come in and say something related to my amount has been deducted without
my permission, so on and so forth. Lots of lots of customer feedbacks, queries be handling it in different ways. All right? So don't worry, you're aware about automation, process mining, nothing. Just think as a layman, as a
working professional, you're getting this problem. I'll give you a minute. Think of a solution. All right? Whatever solution you can think of. Maybe you can say that hey Sri why do we you know uh take data from all of these sources
again and again. Let's make a software in which I press a button in the morning. It automatically collects data from all of these sources, from all of these sources, segregates it, analyzes it and gives
gives me a very beautiful report wherein I can just see the outcome. Could be an learners. Think of what all answers you can think of an example as well while you're explaining. But you got a minute
learners. All the best. I'll just keep quiet for a moment.
thinking learners. But as of now, what we'll do is we'll just try to listen to some of the ideas that you have got over here. Okay. So, let me know if you can here. Okay. So, let me know if you can see my whiteboard. And here we go. Let's
think you want to take this problem up? Anyone? Go ahead. Let's start.
>> Yeah, probably like all these data this could be quite subjective. So maybe like quantify these data >> uh or classify these data and then try to come up with some sort of a trend analysis or uh
maybe uh deviations or uh >> Got it. Got it. So you set some parameters. Less Nik is saying you set some parameters. Let's say on the ratings part or on something related to analyzing what this data really means
analyzing what this data really means and if you see too many negative words horrible poor service so on and so forth so forth that can be highlighted so that this is something you need to focus on right NL somewhat that's what you were
hinting at correct >> uh yes and like maybe like the kind of problems that we are getting or kind of complaints we are getting so maybe >> classify I >> yeah classify quantify [snorts] uh maybe
>> yeah classify quantify [snorts] uh maybe like uh maybe we can try and like assign some weightage to the keywords like five >> five okay so let's say if more and more complaints we are getting or more and
more answers we are getting >> related to customer service all right so that those complaints can be classified as customer service which shows that this is the area we need to work on more all right good nil next learner Next
>> hello >> yes over to you case if you're conducting a survey there need to be a base right for what you're >> you need to have a proper filters for example if it's running a customer
statement >> uh if the problem like is it a customer dependency or is it a system dependency right uh if it is a for example if it's a customer dependency how we can solve it it will be a different al together
>> if it is that we can handle it internally. So it will [laughter] be classified. So I think what the first person said it it will be the same. So understand the journey. >> Got it.
>> Good. So now another similar option it says that you know to work out the feedback or details much better let's have a concept called filters and right Tina and once you have a concept called filters then you can
>> even a good point on this Tina as you mentioned you can put who's in charge if >> sorry in power or in Tableau we get an option
can compare like 30 days 90 days >> yes Yes. >> So we can do that also if we if we get the analysis properly. >> Perfect. So let's say filters you came across a very big problem as you know
Nesh also mentioned it's customer service. All right. Now you can hand over this problem to the right team because as a BA you're working across teams right so now instead of you being the person taking care of all the
problems now you know that this problem has to be handled by Tina this problem has to be handled by Nickel's team this problem has to be handled by Savvitast and this way you are able to narrow down and put accountability
good answers Tina and Nesh anyone else would like to pitch in with any other Hi sir, good evening. Kavia here. >> Yes, Kavia. Over to you. Uh >> sir, actually based on this definitely we are we will be getting multiple
>> Yes. >> So uh in that definitely I could imagine we will get multiple questions on the same uh problem so that we can duplicate
that. Uh >> excellent. So we can make them as a unique those unique uh problems should be compiled and then we can share that with the relevant uh concern team or a person.
>> Excellent answer. So you mean to say another good point K is saying that Sri [snorts] whenever we are doing this kind of an activity we see a lot of duplicates right and these duplicates tend to waste a lot of time. Therefore,
apart from classification and filters, we also identify these duplicates and we remove them so that instead of a mountain to scale, we know what all areas we have to work on narrowly. So, good answer, Kavia. We'll take one more
good answer, Kavia. We'll take one more answer. Anyone would like to pitch in. Savvita, maybe you'd like to go. Snigh, you would like to go. Anyone
>> Go ahead. Go ahead Savvita. >> Yeah [clears throat] I was working for these automation things like um wherever we are finding the um any um issues from
the customer right. So we used to get the requirements and based on the requirements we were like I was working for the clicksense visualization tool where um wherever we have these manual things and wherever we have this issues
what you have mentioned in the query right so that one right so that one >> we were uh improvising that in the means of trends >> showing the trend and the uh showing the
sales also and everything whatever we have the issues. >> Got it. Got it. Learners, thank you. Thank you so much. Uh Savva, learners, let me tell you a very interesting example where this thing was tried and
in one scenario it backfired. Okay. How many of you have traveled recently or maybe you've you know or traveled or anything with Indigo? Just type yes in the chat box. Indigo, Indigo, Indigo customers. Indigo. Anyone?
Go ahead. You can just mention in the chat box. Yes. No. [snorts] Now this thing happened. Thank you Val. Now this thing happened. A customer didn't have a good experience while he was traveling. Okay. He puts up a tweet tagging Indigo
and he says, let's say that customer over here is, let's say that customer is Tina. All right, Tina. So, you went and you had a post on Twitter saying, "Thank you so much, Indigo. I had the most amazing
horrible experience working uh traveling." And uh she put a very nice paragraph sarcastically telling how bad the experience was. All right. Now, now the bot interpreted words like
amazing, best, unimaginable. All of these tended to be entering in a very positive area. And the bot
commented, "Thank you so much for your feedback. This level of encouragement and support motivates us to work better so on and so [gasps] So learners yes these kind of things do help you know uh save a lot of
effort but many a times these things tend to not be as evolved as a human can tend to not be as evolved as a human can be. Okay now 30 seconds you've got 30 seconds go to three go through these objectives and then I'll come back.
Once you have read these objectives, just type yes in the chat box.
We'll be covering as many as real world examples for all of the So, if you have read this learners, just quickly type yes in the chat box. Hurry quickly type yes in the chat box. Hurry up.
uh I'll give you another 30 seconds. Read this then we'll discuss together. Okay. It says use of tech to automate repetitive rule-based task. Hurry up. I'll give you 30 seconds. Introduction to business automation. Now everyone
listen to this very simple example. All right. this session. Snigh, are you there? Snighel.
>> Snhel. Let's say you run a shop. Okay. Sny Any any suggestion? Maybe a book store, jewelry store, anything. >> It will be a food shop. >> Say it again.
>> Fruits. Organic fruits. >> Fruits. Okay. Now learners, we have >> Fruits. Okay. Now learners, we have Snill who has her own
let's say every day she has some customers coming invoices getting generated um and let's say she's also have because she's doing well she has also taken a GST number every day there are multiple invoices being graced
that's how usually a shop works okay [snorts] at the end of the day or maybe at the end of every week day let's say month snake Nahil would like to know
top selling products, total revenue that she has got. Maybe revenue. Maybe she would like to know let's say you know uh which day I had
highest sales so on and so forth. Do you agree that these kind of questions are very common for any business owner? Yes or no? Very common. Yes or no? Go ahead.
She has a name over here. Let's say Snighel Enterprises. Over here she has a GSG number. She has a date. And over here she mentions the name of the product. Over here is the amount.
Over here is the total [clears throat] that the user has to pay. Let's say she would like to know how much sales were done on every day at the
end of the month. Now a manual job would be open these invoices, copy this total, be open these invoices, copy this total, copy this date, put it in an Excel
copy this date, put it in an Excel and then it gives you a table. Right, Snail? That's how a manual job would work for you. Yes or no? Right >> now, if she's running a shop, she would be having almost the same format of a of
an invoice. Yes or no? It is not that she will change the format of her bill she'll do, but she'll keep the same format of the bill. format of the bill. A very simple rule. Copy total,
A very simple rule. Copy total, copy date, paste it in a table and then give me the grand total. Do you think it is a very simple
straightforward logic? Do you agree? Yes or no? Does not require rocket science. I mean even a schoolgoing student can do that. Do you agree yes or no? D you find it confusing. Okay, maybe a high school student can do that.
you agree that this is a very straightforward logic of copying date copying total and putting it in a format and putting in a table kind of a thing in this scenario learners since it's a very simple straightforward rule based
we don't need snighel to waste her time and copy pasting a very simple software can identify this is the date this is the total amount and automatically create a table
for her. Today, this same thing can also be done using chat GPD. However, when chat GPD was not there, there were many other tools who can get this thing done for you. They can open your PDF, they can select whatever you want them to
date, they can select second thing as total and then they can keep on adding adding and adding for you. Do you see that we have saved a lot of Do you see that we have saved a lot of time and effort for snail? Yes or no?
something. Let's say, but do you agree that we saved a lot of time and effort for snail? Just type yes in the chat box.
My next question for you is my next question for you is please take let's question for you is please take let's say 45 seconds to 60 seconds. What you need to do is jot down multiple
multiple other benefits that snail will have when she decides to use a software to get this thing done rather than manually this thing done rather than manually copy pasting data every end of the day.
I'll give you let's say 45 seconds more learners go ahead think of multiple benefits. Maybe you can say that you know now Sneighhel is free from her time of doing this thing. She can invest that time in expanding her business. Maybe
you feel that better insights would help her buying the right products. It's up her buying the right products. It's up to you because now the software is identifying your top products. It is identifying your least popular products.
It is identifying let's say which are your loyal customers because your invoices also have contact numbers. So now she's getting tons and tons of data being analyzed and presented in a beautiful report format for her. And
over here learners let's stick to example that snail had told snail you wrong right >> that's right >> organic fruits all right do you let us remember one thing it's a perishable
remember one thing it's a perishable product right snail so you can think something on supply chain as well that you know she can order her fruits faster so that they not she does not need a coal storage just to ensure that they do
not get what you say spoiled. 30 seconds more learners. After that we'll discuss some of the benefits that Snigh would achieve if she starts using this.
business improvement efficiency uh uh accuracy saving time and efforts. I'll give 10 more seconds learners. Then you need to give me some quantifiable and specific example related. Let's say snhel you have mentioned accuracy less
investment on people work. That's right. Great. Now if it is giving you some accurate answers how is it helping you in what sense? Let's say sury it helps me understand that what time of the year what fruits are being sold maximum so
that I can place orders in advance. This is one example of benefiting from accurate data. Learners, when we talked about Snhel's shop, we had a very straightforward, what do you say? Standardized process. Don't you agree
that this is the invoice, this is the product name, this is the product name, this is the total amount or let's say over here this was the discount that you know she offered. This
is the date and it will remain like this. The and it will remain like this. The software has to fetch only these details and then the software has to put them to understand what trend are we looking at
maybe some questions let's say top selling products worselling products selling products worselling products best sales so on and so forth which are something that snal will use month on month it will not be a scenario that
this month she wants to know okay what is what product is my top selling like to know this. It's something she would like to know on every month- on-month basis. Therefore, do you agree learners that there is a very
straightforward standardized process defined that our friend which is let's say robot we call our software as robot let's say has to follow. Do you agree learners? Yes or no? Just type yes or no that yes is straightforward process.
Copy paste this data analyze and then show it to me in the form of a trend. I will say okay you know why don't you all tell me your names and you know where do where are you joining from do you think this is also a very standardized process
yes or no very yes you're right visual it's kind of a standard operating procedure right therefore when you have these kind of standardizations these kind of standardizations it becomes your first step
it becomes your first step towards business automation although we have about machines today that keep on learning, learning, keep on learning and then they keep on adapting. But we'll not go to that level
in this session. Okay? Because that would take another round of discussion. Now we here with the first step which is some kind of a standardization is needed. All right. Now learners consider this scenario. Today our friend Jatin
this scenario. Today our friend Jatin comes to Snighel and Jatin says hey discussion about your software in session snail says yeah so Jatin says you know what give me 1 cr I'll make this software for
you do you think snighel will agree jatin definitely come and say okay let's go ahead and do that you know let's let me give you 1 cr and you give it to me tomorrow so that I can start using this
software on from a fruit shop. The chances are very less that immediately she'll get convinced that you know this will actually work. Is it even worth the money? So on and so forth. She's very skeptical.
However, Jatin says why are you being skeptical? Yours is a rule-based simple job. There's no questions asked. There is no brainy thing needed as such. Why are you worried? Tell me learners, how should Jatin convince Snhel
that this thing is actually useful for her? Let's say Snighel is again from a non- tech non-software background. Jatan would like to convince Snhel that this looks good. Learners, how can he do that?
Go ahead. Quantify benefits. That's great. You know, show her that you know these all are the benefits that you will be getting if you deploy this software. Great. What else? What else learners? What else you think can do
you know Snehal agrees okay I think uh if we don't worry about the price also if we don't worry about the price also yes this might work let's say 1 cr for snighneighhel is not a big amount she runs reliance
hypermart for her 1 cr software won't make much of a difference but she's has to be convinced that you know this thing will work my question or suggestion for you is do you think it will make sense
If Jatin can give a demo one week of small trial so that Snehiel can understand what benefits will she exactly get from her
benefits will she exactly get from her own system. Do you agree? Yes or no? Learners prototype. That's right. by you know giving her access he prototype app by you know Jatin can say you know you take this app try it for one week or
maybe a month if you feel it's good then come back to me otherwise it's fine come back to me otherwise it's fine therefore learners before we ask snighel therefore learners before we ask snighel to deploy this concept at scale
we need to test it in a very small environment maybe if she's running something like reliance hypermark then maybe she can run it for
one store works good she can go for 10 100 and that's how she can scale therefore learners this is what we were discussing over here that you need some kind of a process standardization you need some
kind of a before you reach this scalable part from process to scalable you need some kind of alignment that this is working maybe your sales team is able to understand your supply chain team is
able to make out some sense from this insights and most importantly they use it because let's face it it's very difficult to change habits if someone is used to excels someone is used to doing
change it they'll say oh wow this looks so beautiful but I will come back to my therefore you need some kind of an alignment within your team so that they accept it and of course you formalized these requirements that I want these
these requirements that I want these kind of insights top products, worst products, highest revenue, date wise revenue, trend analysis, so on and so forth. Now all sorry now all of this that you
see again works around whether the prototype was works around whether the prototype was successful or not. That means when you ran a Jatin ran a prototype with snail's big let's say Reliance Hypermart or
fruit shop did she get the required evaluation successful. So that's how it works learners instead of deploying it to scale you undergo a
of deploying it to scale you undergo a journey and this journey timeline varies If you're talking of something like Reliance, it may take maybe six months But if you talk a very simple uh let's
say store let's say a nearby hypermarket, then this thing can be hypermarket, then this thing can be implemented maybe in one quarter itself. Again uh very simple differences. I'll give you maybe another 20 30 seconds
because this is just acting as a refresher to all the concepts that we are discussing through our examples. All right. Okay, [sighs]
don't need to memorize or something. All it says is what we need to do understand what we are trying to do over here. Let's say we are trying to do some kind. Don't worry about learners of these big big jarens process mining.
Don't worry about these things mentioned here. [snorts] All it says is if Jatan here. [snorts] All it says is if Jatan has to convince Snell because SNHL will pay a very good amount to get this software deployed.
to get this software deployed. Everything has to be documented. Do you Just tell me. Give me a quick heads up. Maybe you can give a thumbs up using a reaction. Maybe you can type yes. Maybe you can speak. Yes. Har. Do you think
that everything has to be documented so that after the deployment of work is that after the deployment of work is done can't say you didn't do this right you agree that everything has to get documented this is where
we see these steps coming in the picture first thing obviously would be Jatin telling Snehil that this is all I can deliver because I've gone through your understanding and what exactly you're looking at so So this is what I can
looking at so So this is what I can deliver. Good. Now before we go ahead sign document and you know go with the payment terms and all let's document it in the form of a very simple document. Okay. This document
Okay. This document will be followed and validated. See this is what I told you I will give and this is what I will give in the form of a software. Okay. This is what it means over here.
First Jatin tells documents and then starts building so documents and then starts building so that this particular thing learners developing a tech product or a software has got infinite scope. Do you agree
learners that today if I ask you to make a software for let's say Reliance Hypermart or for lifestyle or for H&M or for Marks and Spencer's home center you will have infinite number of ideas. Do you agree yes or no? Just tell me yes or
no. If you know we can do this we can get more loyalty points. We can have collaborations with let's say quick commerce. We can so many things could be possible when you talk of developing some kind of technical softares. Yes or
no learners. All right. Therefore at some point you need to limit your scope and this is what this business requirement document does for you. It limits your scope that only this
thing has to be delivered. All right. Are you read this learners? Yes or no? Just quickly type yes in in the chat box if you have read this.
you have recently been? [gasps] So we've got people who have been to decalon right learners. Now listen to this. Let's say
learners this is what you would have definitely seen. There are two types of checkout counters at Decathlon. Can anyone tell what are those two types of checkout counters what Decathlon thought when it understood that you know let's
checkout. So let's say Datathlon said all right you know let's understand if Isa is the person working at the checkout counter what exactly she does step number one she'll come let's say she's there the customer arrives
step number two customer puts up his or her basket all right now once she sorry yeah once the customer has put up a basket then the customer has put up a basket then Isha would pick out each item
scan it and put it away. The software will keep on identifying details
cost this is the total bill this is the item code so on and so forth and [snorts] finally she'll ask isha Isha will ask do you want to have a carry bag or not very common question and then final bill is generated
do you agree that roughly this is the process that is followed at any checkout an example over here. Do you agree learners? Yes or no? >> Yes. >> Right now decathlon realize that we've
>> Right now decathlon realize that we've got so many steps over here. Let us see if something can be done in the automation part so that this entire thing can be done using a selfch checkout.
customer arrives great that's step number one. Step [snorts] number two, number one. Step [snorts] number two, let customer put items on his or her own in the basket. Step number three, the
basket should analyze using an RFID scanner or a tag. This is the item. This is the details of the item. This is the price. Add it in the what do you say
screen so that the user can see this item has been added. Step number four, the user keeps on adding items. Step number five, there are some safety checks, right? That hey, you know, uh the user should not
let's say add an item and just take it out. So let's say some kind of a safety checks are present and finally the billing statement and all is done. Therefore learners when people were developing this product they analyzed
developing this product they analyzed these steps which is a part of your checkout process and that's when they created a cell
checkout process. Therefore learners this process that you know I would like to understand first what is being done manually how it is being done manually create an automation software or something for that it is called process
something for that it is called process mining being done now in simple words we call it asis
it asis and where do you exactly want to be and where do you exactly want to be that can be your 2B over here. All right, this is very simple example of how process mining works. Now this
of how process mining works. Now this same thing can be applied to any process where you feel some kind of an automation could be a possibility. understand what is process mining just quickly type yes in the chat box. And
understand the meaning of this jargon hurry up just drop me a yes in the chat box so I can go to the next one because I want you to understand where these things are applied and uh that's when you'll be able to understand better
when you'll be able to understand better okay reading I'll just share one example keep on reading but listen to this example
Novatis. Can anyone answer? Noatis works in which field? Which sector? Noatis. Now it's a it's a pharma company. Let's say they come to innovation hub of Accenture in Bangalore. We call it BDC1.
All right. Now what happens when they come? We've got multiple what do you say tech related ideas that we would like to share with them. In simple words, it's
just a story. Now what happens? They say that okay, we find this particular story interesting. Let's do a P or a POV. You know, let's just test it whether it it's going to work or not. Okay. So, we do a very
simple small scale trial run. The cost is equally shared by the vendor and the company and based on the feedback we get then we scale it to different areas. So this PC or POV nickel over here is acting kind of a validation that this
acting kind of a validation that this will work yes or no. And and since usually the cost is shared between Accenture and the client the risk factor is equal for both. Therefore the client is also not very
hesitant in putting some amount in testing this. So that's how usually a validation works. All right. This is again steps to process mining. You understand? You prepared data so on and so forth. We've already discussed this.
We'll come now. We'll come now and take a pause on this. There is that's where your soft skills come in the picture where you put
something called as a full stop bus. Now we have done enough otherwise if you keep doing this it will keep on changing again and again. See when we were working with clients from Middle East, we would realize that there would
be frequent changes in the leadership as in with whom we were supposed to work. All right? And anytime a new leader would come, they would change the entire scope again. This was a trend which we had observed
from the past. Therefore, every time this would happen, we were well aware that this would happen. However, the clients were high-paying clients. So the company doesn't bother if let's say today Issha says I want to get this ABC
product done and then Jatin and Snehil take over in the next phase they say that okay okay what you have done just leave it where it is I want you to go XY Z okay you go XY Z it's fine therefore isha it depends on the clientele that
you are working with these all are how things work when you work go ahead with the client all right Isha so yes it has to be a soft skill middle ground that to be a soft skill middle ground that you would be playing. Okay.
you 10 uh maybe 30 seconds. Read this quick learners. It highlights project objectives, what's required to accomplish the project, so on and so accomplish the project, so on and so forth. Hurry up. Read this.
the BRD? See it now there's it's not exclusively mentioned that the risk part is over here working out like this however
here working out like this however however when you pitch these kind of let's say I'm pitching this kind of an automation product to you this is something you are experiencing and that's coming out from your end so
that's when you come out how do you mitigate that risk and that comes as a mitigate that risk and that comes as a part of your BRD word no we identify the risk and we identify the process or steps that we
identify the process or steps that we are taking for risk mitigation. You getting somewhat an answer to this Tina? Now learners uh it ensures that business needs are clearly defined communicated so on and so forth. Uhhuh.
read this quick because I really want you to open chat GBD as soon as possible. you know explain a business requirement and then let's see what kind of a BRD we are getting learners always remember one thing
remember one thing formats templates how it is written what all components would be there this all is highly dependent on the client and the vendor that you are let's say you're an IT services vendor's own and so forth
so there's no what do you call a constitutional law that this is the document for BRD. Okay. Therefore, I just want you to Okay. Therefore, I just want you to experience how do you create a BRD?
I want you to quickly go through these steps. Then we'll identify a business problem kind of something and then we'll create a BRD. Okay. screenshot? Have you understood? Yes or no? Just take the screenshot because
these components we'll be putting in chat GPD and we'll be creating our first BRD now. Okay. So I want you to take the screenshot. Be ready. We'll be taking up a problem statement together. And once we have taken the problem statement,
we'll work together to create a sample BD. All right. And if possible, we'll be creating two such examples today. Okay? Hurry up, hurry up. Just drop me a yes in the chat box. Quick, quick, quick, quick. Hurry up learners. Just drop me a
yes in the chat box so that I can proceed.
Thanks, Nikil. Hurry up learners. Just open chart GPT and listen to me. Remember learners, we discussed something related to our friend Snigh's
something related to our friend Snigh's fruit shop. Yes or no? Do you remember? fruit shop. Yes or no? Do you remember? Now you can consider Snigh's food shop as an example. Fruit shop. You can consider Tina's boutique as an example.
You can consider Decathlon store as an example. Okay, [snorts] remember we were creating some kind of an analysis or an analytics tool for Snill to analyze her top selling product so on and so forth. Yes or no? That's what was the problem
statement we worked. Do you remember learners? Yes or no? Hurry up. Just let me know. >> Now learners, we will put exactly the same problem statement over here. Okay. How do we do that? You just click on
this mic option over here and start speaking that we need to create a business requirement document. This business requirement document we are business requirement document we are creating for an organic fruit retailer.
So let's say they have got 10 stores spanning across Delhi NCR. They deal in only organic foods and they have been in the business for the past 5 years. Just [snorts] Now the business is looking forward to
some kind of an automation tool. This automation tool should analyze sales data. In this way they can make better decisions pertaining to procurement, supply chain, customer satisfaction,
loyalty plan, so on and so forth. Okay. Five, seven examples. Okay. This is what we need to do. Now once you have done that I ask you to take a screenshot of the components of a BRD. Put that over here and then ask
BRD. Put that over here and then ask chat GPD to create a BRD. Now all right you can take any business give as many assumptions as possible and uh just ask RGB to do it. Just in case you're feeling confused or
something you may not be able to follow on how to do that. We'll make it very follow me on the screen. All right, I'll be also doing the same thing so that you can follow me on the screen as well. If you think you have understood what has
to be done, you're more than welcome to carry it on your own way. But if you feel that you know I need a little help, so just follow me step by step. Are we good learners? Can we start creating our first BRD? If yes, hurry up. I want
Hurry up, hurry up, hurry, hurry up. [snorts] So that I can now start and uh own, go on your own. You can create a BRD for the boutique for organic food
retailer, it's up to you. I'll be doing it for snails organic fruits and [snorts] uh you can follow me if you want. All right. All right. So let's requirements over here. I'll just speak it out instead of typing and all. All
right. We need to create a BRD business requirement document. This is for an organic fruit retailer based out of Delhi NCR. They have expanded to Bangalore, Hyderabad, Pune, Ahmedabad, Gandhiaga, Mumbai, Chennai [snorts] and
Gandhiaga, Mumbai, Chennai [snorts] and they've got over 100 outlets spanning across these cities. They deal only in organic food, organic fruits. >> They have sales data spanning across two years, 3 years and regular customer
footfall takes place. Now they would like to make what do you call a business automation software that would analyze all invoices to generate insights from sales data so that the business owner can make choice make informed decisions
pertaining to specific areas of supply chain customer satisfaction loyalty plan uh procurement opening of new outlets uh just ensuring that you know the uh the
procurement does not go waste. Does she need a coal storage or not? Where all she can expand further? What all new product lines she can open? Therefore, she needs insights on of all of these areas. We need to create a BRD for this.
The format or the components of the BRD are attached as screenshot. This is one few assumptions learners that I have given over here. You can give it yours. It's up to you.
I'll just attach the screenshot over here.
what happens. What is a BRD that I am able to get. Okay, all [snorts] learners who are able to get some kind of an output, just drop me a yes in the chat box. Just write your name in the chat box. Okay, so that I can understand how
box. Okay, so that I can understand how many of you are up to speed ask you a very simple question on this. Okay, very simple, very straightforward.
Okay, now since you're doing all of this, so just listen to me for a moment with whatever remaining things you've got to complete. Okay, can you see me? I'm sorry. I just got tired of sitting. Hi, can you see me? Yes or no? Yes or
Hi, can you see me? Yes or no? Yes or no? Okay, learners, what we just saw was that it is so easy to create a BRD. Do you agree? Yes or no? All you need to do is just copy this entire thing that I gave you. Put in some of your additional
Even if you don't do that then also it's fine. Just paste it in charge GPD and press enter. Your job's done. Do you agree? Yes or no? Right. >> Absolutely.
>> Then then the question arises where does your role come in the picture >> to put the correct prompt. >> Yes. This is one you need to put the correct prompt. Great. The other place where you can work as a business analyst
is no client remember this learners when you sit face to face with a client no client has got the time to go through this huge document that's what happens in when you're discussing with your
client he'll or she will not go through the entire whatever chart GPD or let's say no whatever document you have got they are only interested going to do tomorrow when it will be delivered
What is the expected outcome? If it works well, then we go to the next phase. Then we go to the next phase. Therefore, learners, you let the BRD be Therefore, learners, you let the BRD be what it is. Okay. Most important concept
comes up is once this entire scope has been identified, once client has been told this is what you will be delivered and this is what is there in BRD, then comes the role of a BA. Are you able to break it in phases to
get it delivered? So that step comes after your BRD has been made because let's face it when someone put it for decathlon when someone put it for Tina's boutique or for Nida Nida's let's say organic fruit
retailer you know this entire document it's a forever job I mean you keep on adding things you keep on adding things it will just go on and on. Therefore the BA role comes in the picture on how well you're
able to structure it in such a way that you're able to identify from this BRD snighel we'll deliver this thing tomorrow from this BRD we'll deliver this thing in a week's time and this is how the phase deployment will work so
scope is this but within [snorts] scope you must realize that everything can't be delivered in a single day so this is the timeline that we are looking at This is where the role of a business analyst comes in in communicating how this will
get delivered. Chat GPT and all tools can help you break down the scope in phases. But to get it across the table, take feedback and then rebuild it. That's where the BA's role comes in the picture. BAS today, product managers
today don't need to sit and type things. Those days are gone. Therefore learners you need to use whatever answer you have got to take feedback from the client divide it in phases and then start delivering.
So if you're clear with this entire story and all what we did till now, just give me a quick heads up. Yes, in the chat box har you know we are able to understand that this is what you were hinting at and
this is how we'll be able to deliver it. Till now anyone has any questions with whatever we have discussed you're more than welcome to ask now because many of
you tried it for the first time and uh there would be certain doubts you have. So you uh you know go ahead tell me any questions any concerns any examples you would like to discuss because we'll take the session at your pace. Go ahead tell
the session at your pace. Go ahead tell me. like this uh business requirement document it still uh is in the form of
bullet points or a very high level maybe 50,000 view >> but when it comes to the detailing of that like let's say if I mention in the BRD that I have like a requirement for
like maybe a sales report a profit report and inventory management report but then how that uh report is going I mean where do we source the data columns should be there. I mean those nitty-g gritties and detailing of the
things. So will that be a part of your BRD or that will move to the FRDS and other documents? >> Okay, good. Now learners, let's have a >> Okay, good. Now learners, let's have a look at a realistic scenario that this
report would be able to give you this data. This particular part of the report data. This particular part of the report will analyze this way. However, we are will analyze this way. However, we are not getting in details
like right Nikl or from where you're planning to source the data and most importantly let's say I need to understand first of let's say I need to understand first of all exactly what the report will show
I don't want to hear as Nikl mentioned from 50,000 ft height that's right that's what your question was Nikl >> yes correct So learners there are two ways on how this can be
worked about. How I have worked in my career. Once we finalize these kind of requirements we move to a document called PRD. Anyone aware of this? Anyone
any guesses what exactly this is or something? Anyone? >> Process uh requirement document. >> Close enough. So it's simply called a >> Close enough. So it's simply called a product requirements document.
product requirements document. Okay. Now a PD is something learners Okay. Now a PD is something learners that goes in every nitric detail that Nikl was mentioning. That will first of all ask hey you know what you mentioned
something about a report. That's great. Please tell me what all exact components the report will have. If you it will have these components. Let me know the sources of data data that you are looking at. If these are
the sources of data, let me know the process of extracting them. That's where the PRD comes in the picture and if you are making let's say one particular generation kind of a thing and it caters to multiple different domains.
to multiple different domains. Therefore the PR also just like your BD is an ever evolving document. It will cover variety of aspects of your cover variety of aspects of your products and every time you let's say
products and every time you let's say break down BRD into four phases. So this will be executed by a separate phase of your product. This phase will be executed by a separate phase of your product. So on and so forth. So the
detailing part happens usually in the case of PRD which is your product requirement document. If [snorts] you're developing a product in the scenario of a chatbot or the insights report we are essentially developing a product. So it
narrows down to the PRD. This is where all of those details are mentioned for any product that you are trying to develop. This is one way of working. There could be other ways in which BRD only goes on to emphasize and there
could be other contracts but as a BA or as a product manager you work across these two documents. So Nicl this is what I have worked maybe there could be other example areas in your service in your areas that you would have worked
but this is one example how things have worked with me. Finally we are here wherein let's create a very simple example of RPA. Okay, I'll tell you how
example of RPA. Okay, I'll tell you how this is to be made. But but quickly have a look at this. Just have a look at this and we'll discuss now because now let us create a very simple example of RPA. All right. And that very
simple example works this way. I've prepared some uh what do you say documents for this also that you can use. Now the simple role over here is use. Now the simple role over here is today we have a bookshop. Okay. Let's
today we have a bookshop. Okay. Let's say Nikl has a bookshop and Nikl's very simple requirement is I should be uploading my invoices and the RPA or whatever tool you have made be it you know uh it's called a GPT
made it should pick out date pick out amount and give me day wise sales data that's it something what we
discussed for snail so we'll be creating that now in the form of a GPT. All right. Therefore, I want you to be with me. The GPD or in very simple terms, understand it as a bot, as an agent, as a bot, whatever you would like
agent, as a bot, whatever you would like to call it. What this does is it needs a to call it. What this does is it needs a PDF of an invoice. After taking that PDF invoice, it will pick up the date, pick up the amount, put it in a table. You
put next invoice, update the table. You put next invoice, update the table. This put next invoice, update the table. This is what our robot or GPT would do. This is what we'll be creating together. Now learners, this is a very simple example.
learners, this is a very simple example. It can get even complex. It can also say apart from this table, I also want to know the top selling products. That thing also you can add. But for simplicity and launching the first
version, we'll do it now with very simple analyzing of date and amount. If you're clear with the problem statement, type yes in the chat box and let's start. Go ahead. Go ahead. Type yes in the chat box quickly because we'll be
using something called chat GPD because it's most commonly accessible. Let's come back to the homepage of chat GPD. Okay.
have this sidebar you'll have an option to open sidebar so open sidebar you will have an option called explore GPDS all right if you can see this option over right if you can see this option over here I want you to click on this and you
will see a page that looks like this first we'll do this first what we'll do first we'll do this first what we'll do is we'll see how this process works then then what I'll do is once we have created some of us have created this GPT
we will share this GPT's link so that you can try this GPD on your own creating this GPT is a 2 minutes task which you can create after you have left the session you can have a look at the recording because that's not the scope
because I want to give you more hands-on experience okay therefore it's a two-step process of creating a GPT however All of us can use that GPD no
matter we have the pro version or not. All right. So if you're good with this approach learners quickly type yes in the chat box. Hurry up that okay you can being created because that's the closest we can get on you know having our own
agent being created today's session. All right here we go learners. So all we need to do is those who have access please follow me. You have an option called plus create over here. Just click on this and just mention let's say this.
Look at my screen. What would you like to make? So I would What would you like to make? So I would like to make a GSG invoice analyzer. Okay, just type this everyone. Those who
have access, those who don't have access, just bear with me and look at my access, just bear with me and look at my screen. Okay. all. Would you like to change the name? No. No. This name looks good.
No. No. This name looks good. Okay, that's it. will try to create a profile picture or something of our uh GPD.
something of our uh GPD. Let it do that. is told what changes you would like. No changes. This is good. So, let's not
waste time in changing the images and the name. Let's focus more on what we would like this thing to do. All right. Now, it now it says what should this GPT prioritize? You know, compliance checks or what should it do?
Now what we'll do is we'll elaborate. We say that user will upload say that user will upload invoices. Okay, which will have
invoices. Okay, which will have date and total amount. I'll just Okay, date and total amount. I'll just Okay, total amount. I want GPT to fetch
total amount. I want GPT to fetch date and amount. total total amount date and amount. total total amount and create a table date wise. and create a table date wise. It should show
Nothing else. All you need to do is press enter. Now All you need to do is press enter. Now let GBD do its job.
focused on extracting dates, total grouping and then giving you date and total sales. No extra GST analysis commentary unless requested. And it also
says you can try it in the preview panel by uploading a few invoices. Okay, by uploading a few invoices. Okay, upload this invoice over here.
you to gives you total sales on that date. Click on create. Let it load for a moment
Okay everyone, click on create. If it doesn't work, click it again and click doesn't work, click it again and click on GPT store and click on save. Do this learners click on GPT store and click on save.
Everyone if you have opened this screen type yes in the chat box so that I know how many of us who were not having access to creation of GPS at least can play around with the automation agents or robots that we are making. Okay,
here, hurry up, hurry up. Just keep typing. Yes. Thanks Tamara. click on the hide option over here. The next the option next to stop sharing.
Yes. Now, Tina, you have an option of plus. Click on the plus option and upload all invoices, all the PDFs that I shared earlier, right? So, yes, all of these. Upload all of these. Now select all of them and just press enter. Not
all of them and just press enter. Not the C. Yes. Press end. Open. Good. Okay. Five files at a time. So not using just upload any first five. Okay. just upload any first five. Okay. [snorts]
Yes. The five one. The five I think you have taken six. Yes. Okay. Now just press enter. Let it load. Let it load. Let it load. Yes. it load. Let it load. Yes. So you see now you've got saleswise data
and no no don't just don't click just click on cross you see learners how we created a very small automation bot or an agent for our own self without an agent for our own self without spending any coding resources time just
by using a few very simple English prompts all right showc learners let's quickly go it says it's a trigger it's a board that defines steps it's a workflow. Please do not worry about reading, understanding, memorizing or
anything. This is what you have done now without going through the entire theoretical process. You have already created your own RPA today. All right. These are the characteristics. It follows rules uh you know changes let's
say unless until the concept of GST only vanishes that's when you would need to to expand the scope of the project that's when you want to make changes to the GPD. That's what it's highlighted as a characteristics of RBA.
>> And if you scroll down, you will see this course business analysis capstone. this course business analysis capstone. If you go inside this um what you will see is there are under the assignments section there are three different
projects. So the first one is credit card for fraud analytics. Let's take a look at what the project is. Um the objective of the project is to identify unusual credit card spending patterns and fraud risks by creating a
patterns and fraud risks by creating a BRD and using SQL, Excel, PowerBI enabling datadriven decisions for financial compliance and fraud mitigation. So what does this mean? So uh I am assuming uh all of you are
familiar with credit card but essentially these are uh paying instruments that are issued by banks to individuals and in some cases even to corporates. Uh the way it works is just like a debit card except that you don't
need to have any balance in your account. It is a it is a credit as the name suggests. Uh the bank extends a credit which is um based on your uh credit history and whatnot. uh they will provide your credit uh in your credit
card. So let's say you have a credit $5,000. So up to $5,000 you can use your credit card to make payments. Now the problem is sometimes the credit card problem is sometimes the credit card details gets stolen or um they call it
skimming. So they skim it some uh you know nefarious elements in the society. When you use it in some locations they steal your credit card information and replicate that and they use it for making fraud fraudulent transactions.
Okay. So all credit card companies, all payment organizations will have some sort of some form of uh fraud uh prevention um um rules and regulations
or fraud prevention approaches. Now the specific one that you are going to look at here is you will be given a data set. Uh the data set is basically if you if you can imagine a credit card uh issuer or a credit card network provider. So
the issuer is like the bank. Um so if you take globally banks like HSBC you take globally banks like HSBC um uh standard charted uh city uh Chase these are all uh examples of banks that issue credit cards. Uh so they are the
issuers. Now there are some network providers site like Visa, Mastercard etc. And then there is also uh MX which is both issuer and um uh network is both issuer and um uh network provider. Uh so these uh issuers
basically the issuing banks uh would essentially be supporting uh a large number of transactions at a at a given point in time. Now what happens is that for each transaction they will look for certain flags that will tell them that
certain flags that will tell them that this is possibly a fraud transaction and this is possibly a fraud transaction and then they look at uh past data which is in the last 6 months let's say there were 1 million transactions how many of
were 1 million transactions how many of them were uh fraudulent right so the them were uh fraudulent right so the goal of the project is to find out a way goal of the project is to find out a way in which we can identify um uh fraud or
in which we can identify um uh fraud or reduce fraud by um analyzing the spending patterns. Now there is an organization analyst IQ Edge they look uh at helping financial institutions transform transaction data so that they
can use the transaction data to make decisions. decisions. uh the focus is on making datab uh the focus is on making datab decisions. Um there is rising digital
fraud. We already spoke about that. Uh the BAS will have to identify hidden patterns, evaluate unusual transaction behavior and deliver insights. Uh so that's the uh idea. Uh you are responsible for examining the credit
card transaction data so that you can identify patterns that may indicate fraudulent behavior. Uh so you you need to create a BRD which is a business requirements document. You should have been taught about a BRD in your um in
your CBA uh uh session or the CPAP course. Uh so the BRD will be the place where you will define the project objectives stakeholders cost benefit analysis. [snorts] It will be the foundation for your project. You will
also use tools like SQL, Excel, and PowerBI with some AI features in order PowerBI with some AI features in order to uh look at the transactions data and to uh look at the transactions data and detect any problems in spending behavior
uh based on things like geographic region, segment, uh fraud, occurrences by category, job, and demographic attributes. attributes. Okay. Now let's take a look at the data
Okay. Now let's take a look at the data that has been um given to us. So there are two data files that you will see in your um for your first project the credit card uh transaction fraud check process uh project. So the first is the
credit card uh transaction data. So if you look at this particular data set there are approximately 42,000 transactions that has been captured. So for each transaction um there is an index which is a unique
number. Uh there is a transaction date and time as to when uh the transaction has happened. The credit card number. So you have uh the credit card number that
was used in the transaction. The details of the merchant that uh did The category of the merchant. So for example, this is shopping in online
example, this is shopping in online shopping. This is shopping in a um in a POSOS which is a point of sale which is the actual machine where you go and uh the actual machine where you go and uh swipe it. Um so there are a number of
different categories that you can see here. There is the amount of the transaction, the first and last name of the person who is the owner of the credit card or the card holder, the gender of the uh card holder, the
address in terms of the street, uh the number, the city, state, the zip code. Now you can see the latitude and the longitude of the particular address. So
you you know where the uh transaction is taking place, the population in the taking place, the population in the city, the job of the um of the card holder. Uh then you have the date of birth of
the card holder and there is a unique transaction number and then there is a transaction number and then there is a flag which says is fraud. So uh if you flag which says is fraud. So uh if you filter on is fraud you will see some of
these transactions are flagged as one. So there are around 180 out of 42,000 So there are around 180 out of 42,000 which has been flagged as um which has been flagged as um sorry
as uh fraud. So you can see this. So this is the uh data this is the first this is the uh data this is the first set of data that you have been provided. Um the second set of data that you have been provided is the location data. So
been provided is the location data. So for each credit card so this is a smaller data set which is only around 980 979 uh to be more accurate. And for each of
them there is a latitude and a longitude which is basically an address that has been given. Okay. So this is the data that you have been given. Now let's take a look at the questions that you need to address or
questions that you need to address or the uh what you need to uh solve for and what you need to submit. Um if you look at the uh the um assign assessment
section and you go to the credit card fraud analytics, you will see you can download the complete problem statement here. This is where you will get a couple of PDF files which will have all the uh problem statement information and
then then under the reference material is where you will get the uh the Excel files that we looked at. So you can see there is one for each of the uh projects but the one which we saw for credit card analytics is the one that you have in
analytics is the one that you have in this particular link. Now the uh the problem statement you will see two files. So the first one is a very highlevel file which talks about the problems and what are the tasks. Okay.
So basically what they are looking at is we looked at what is the objective, we looked at the description which is basically exactly what we just saw and basically exactly what we just saw and they will give you uh details of the
problem statement which is essentially the same. Um and then what you need to do, you need to do uh business requirements document. You need to requirements document. You need to create data exploration, pivot and as AI
create data exploration, pivot and as AI assisted summaries for using MSXL. Uh use structured querying using SQL. Use PowerBI to create dashboards PowerBI to create dashboards and use some uh data visualization
techniques and use some um descriptive and diagnostic uh analytics to understand the past behaviors and identify anomalies that may indicate fraudulent activity. The tasks are there. We'll come to the tasks in a
minute. The output dashboards will look like this. We'll see the details of it in a minute that need to be created in PowerBI and in Excel depending on the uh on the question. So there are essentially seven uh tasks for each of
that there is a reference that has been given. You can refer to these um for you to understand how to do this. Okay, let me go through this again. This is basically a high level the capstone project. Um the objective and the
description same as what we just went through. Um then the problem statement through. Um then the problem statement uh the uh relevance in terms of what are the tools that you are going to use. The tasks we will come to the tasks in a
tasks we will come to the tasks in a minute. the output screenshots which is uh basically what you need to attach in your um in your submission uh and the references that you can use for each of the tasks.
introduction problem statement etc which is actually quite straightforward. So is actually quite straightforward. So the first is the BRD right. So you need to document the business requirements using a business requirements document.
using a business requirements document. Now they have given you a template which Now they have given you a template which has essentially uh a set of um sections. So the sections are you need to have a project header. So you can provide the
name as uh credit card fraud detection dashboard or whatever right. uh you can use the current date for submission uh and and mark the document status as
draft. You need to provide a highle summary using whatever you have read. Um you need to provide a list of objective. It should be very clear. Um I think some
of you might remember smart. Um hopefully all of you remember what smart is. So your objectives try to keep it as smart as much as possible. smart as much as possible. Um then uh you you explain the scope in
scope, what is in scope, what is out scope. Um what are the business requirements? What needs to be delivered in order to meet the requirements? Identify the stakeholders who are involved in it.
involved in it. uh any constraints, risks, um any limitations, uh all those kind of information uh needs to be highlighted in the BRD. So, BRD is the first deliverable, right? You
can you have been given a link to a document template, you can use that. You can refer to your CBA course for reference as to how to create a BRD. So, reference as to how to create a BRD. So, that is the first uh task. Uh the second
that is the first uh task. Uh the second set of tasks is on Excel. The first is set of tasks is on Excel. The first is data exploration. Right? So on the in the in the data set there is two u data elements. One is amount and then
there is a city population. So you need to create some statistical summary on to create some statistical summary on this um uh on amount and city population. Basically things like how many values are there for amount and
city pop. What is the highest amount? What is the lowest amount? Minimum, maximum or highest, lowest? What is the average? What is the mean? What is the median? So average and mean are the same. So what is the mean and what is
the median? What is the standard deviation of um the amount? Um all those kind of um statistical information you create as a summary uh
amount. Same for the city population as well. What is the highest population? What is the lowest population? so on and so forth. Okay. [snorts] Then you need to plot a histogram on the amount. What is a
histogram? Histogram is kind of like a bar chart. Basically, it will have these various uh elements. So you can create that on the amount. So it can be by um
category, it can be by gender. So you need to provide those assumptions called need to provide those assumptions called out in your submission. uh you need to provide the number of frauds right so you need to look at so
we saw there are a total fraud of 180 right you need to show that by gender so how many are there where the customer is female and how many are there where the customer is male and by category so if you look at if you looked at the data
there are there is a field called the category so for in that category you need to show for each category how many frauds were there in the data set. You also need to show the
top three states. Right? So in the data set you will see a field which is your set you will see a field which is your um your state. Okay? Now it's not it's not actually a state. It is it is it is a twolet code, right? So it's for
a twolet code, right? So it's for example it's um NH okay IIL. IIL I think example it's um NH okay IIL. IIL I think is for Illinois uh TN for Tennessee so on and so forth right it's it's basically all these are um all these are
basically all these are um all these are um representatives of like twoletter representations of states so but you need to identify those states that has need to identify those states that has the highest number of frauds okay
the highest number of frauds okay top three states then so that's the the second task the third is data analysis in Excel, right? So, you can use the data analysis tool pack which is an this
is an addin. You should know how to add this and you can also use AI feature in order to identify the transaction data. data. So, you need to look at the relationship
city population. What is the relationship? Is it is there a relationship? Right? Look at the fraudulent transaction in each category using pivot tables. Hopefully you know
how to create a pivot table and then by category you need to identify the fraudulent transaction. Uh create a report showing the average transaction amount by job ro. Right? So there is a field in the data set which is a job ro.
So you create a pivot table and a pivot chart which will show the average chart which will show the average transaction amount for each job role. transaction amount for each job role. Um then you can use AI.table to generate
AIdriven summary of the fraud risk categories. What are the categories categories. What are the categories where you get uh the highest kind of um where you get uh the highest kind of um AI risks? Uh you can use chart GPT4
Excel. uh there is a you can add again that is an addin that you can um add into your Excel for highlighting the top risk risk categories and any any
possible mitigation actions right so you can use chat GPT quite liberally right uh the goal is you for you to understand how to use it effectively so that you
can do this much faster than you would have in a traditional fashion so those are the tasks for Excel. Um a few charts, few um diagrams, few tables,
etc. Then there are the SQL tasks, right? So in SQL task, you need to create a schema. You need to provide a name to the schema. It should be your name to the schema. It should be your default schema and then create a data
tables. Create tables. The first table is with the credit card data, the transaction data, the 42,000 plus uh transaction that we looked at. The second is the location data, which is like the 979
um um credit card information that we have with location, right? Um and then have with location, right? Um and then you need to do a number of queries in you need to do a number of queries in the SQL database to do the following.
transactions in the credit card data table. Second is identify the top 10 table. Second is identify the top 10 most frequent merchants in the CC data
do with the fraud. This is like the the most the most frequent merchants. The average transaction amount for each category using the data table.
number of fraudulent transactions and the percentage of transactions that they the percentage of transactions that they represent. So this is another u uh query represent. So this is another u uh query join hopefully you have learned joins uh
join hopefully you have learned joins uh join the two tables so that you can identify the latitude and longitude of each transaction. Uh identify the city with the highest population in the location data table.
find the earliest and latest transaction dates in the CC data table and then you need to do some more data aggregation. So you need to do some
totals. What is the total amount spent across all transactions on the CC data table? How many transactions occurred in each category? What is the average transaction amount for each gender, male and female?
Which day of the week has the highest average transaction amount in the table? So those are the questions from SQL. So with that you will complete Excel and
with that you will complete Excel and SQL. Okay, we'll [clears throat] carry on.
So that concludes SQL tasks. Then you have the PowerBI tasks, right? Then you have the PowerBI tasks, right? So you need to then um create a
uh so that you can understand the spending patterns and fraud distribution. So the first is show the amount spent by different genders on different categories through a box and a whisker plot. Now they have given more
detailed instructions on how to do that. how to um you know you need to open PowerBI um get uh extract the text data uh
create the visualization collect um search for box and whisker chart uh click on the so they have given the details on how to do it right so I'm going to not going to read through that [snorts]
excuse me you are given the latitude and longitude so you need to show how uh the how is the geographical distribution of the transaction. So you can see again
the transaction. So you can see again the steps for that right. Um fraudulent transactions right so you can see which in a map geographically which
part of the world does the fraudulent transactions come from. visualizations. So the box and whisker, transaction map, fraud map, all those
transaction map, fraud map, all those three you add into the visualization. Make sure that it is clearly laid out etc. Um use quick measure a assisted quick measure to calculate a number of things.
Uh the average transaction amount per category. Again steps are here. visualize the autogenerated DAX measure in a column chart. So these are the in a column chart. So these are the steps that you need to do in PowerBI.
Now the last part that I will talk to before I take any questions is how do before I take any questions is how do you submit this right? So once you complete all the tasks what you do is just like in the just like we discussed
in our previous uh session you create a in our previous uh session you create a word document where you will um input word document where you will um input everything. So the first will be
so the first will be the BRD sorry things. So so so the all the details of the BD let's say it runs to a number of the BD let's say it runs to a number of pages then for Excel you will have
pages then for Excel you will have basically task two you copy paste any um pivot tables pivot chart any histograms etc that you have done and then again on
etc that you have done and then again on um on the uh SQL um on the uh SQL um you can provide um either screenshots um you can provide um either screenshots from your SQL uh uh pro query dashboard
from your SQL uh uh pro query dashboard or um you can um you can take screenshots and attach it here. So that is SQL part. So we provide task is SQL part. So we provide task four and five I think and then PowerBI
again um screenshots from PowerBI for task six and task seven. You save this task six and task seven. You save this and then in the um
in the in in the assessment section.
In this assessment section there is a submit option where in add uh for write up screenshots and source code for all three you can uh provide the same inputs
and then submit the project. How this type of projects are done in real time type of projects are done in real time like in companies. Um [snorts] so in companies what would happen is that uh you would have uh done uh you will have
you would have uh done uh you will have some sort of a um analytical model already in place. What you keep doing is testing that against new information on an ongoing basis. Right? So you're uh um so this is used for a
number of things. Right? So you are basically um analyzing past data and basically um analyzing past data and then you flag transactions that are um that are fraudulent that you suspect as f fraudulent and then there is like the
f fraudulent and then there is like the the the the current trend is to try to the the the current trend is to try to do this on a realtime basis where you do this on a realtime basis where you get a notification when certain when um
so what they do is they will take these past um transactions. They will run it through uh analytical systems or analytical engines and then they come up with a model which will flag any new
with a model which will flag any new transaction which is likely to be fraud. And when a transaction which is likely to be fraud, they will then do a confirmation or a check with the customer. Hey, I I see we did a
transaction. According to our system, this is an unusual transaction. Can you confirm that you did this transaction? So, that will help them catch uh fraudulent transactions quicker. So, as a BA, what you should be comfortable
a BA, what you should be comfortable with is again 42,000 is a very small uh amount of data essentially. Uh the data will be significantly more. So things will be significantly more. So things like SQL and even um non uh SQL no query
kind of uh techniques you should be very comfortable with where you can analyze the data. You should be able to uh understand the domain very well in the to achieve. What is the goal of doing all these things? So you will see
different objectives like this is basically fraud reduction. Um in retail basically fraud reduction. Um in retail it is more around uh sales. How can we increase more sales, more effective sales in uh healthcare? It is around uh
the well-being of the patient. So the the objectives will vary. So you need to understand that domain. You need to understand the data. You need to be able understand the data. You need to be able to analyze that data and then provide
u you know insights to the business who can then make decisions based on the insights that you provide. So that's how the projects are done in real time. Um it will be much more intensive but
essentially the the conceptually it is exactly the same. Okay. So patient readmission analytics identifying high-risk patients for proactive care.
high-risk patients for proactive care. So to uh the objective is to identify patient uh quick check. Can you guys hear me and uh quick check. Can you guys hear me and see my screen?
so the objective is to identify patient readmission risks and care trends. So the the tasks remain the same. You need to create a BRB. You need to analyze to create a BRB. You need to analyze some data using SQL, Excel and PowerBI.
some data using SQL, Excel and PowerBI. Uh so the goal here is to support healthcare resource optimization. Um now readmit guard analytics is a BI business intelligence solution provider for healthcare like previously they the
domain was um credit card but here it is healthcare. They specialize in the healthcare. They specialize in the patient data analysis in order to predict the the likelihood of readmission and accordingly plan
Uh so there are a number of tools that are available to you in order to enhance the quality of care uh patient outcomes and reduce the readmission dates.
uh this is important because you know the patient outcomes needs to be better the patient outcomes needs to be better with lower operational costs. Okay. So and the goal here is to create actionable uh insights.
actionable uh insights. So we will um again we will have a data So we will um again we will have a data set. uh the goal is to look at the readmission of the patients and see how we can predict or understand the factors
that lead to the readmission of patients. Now uh the data set here
um I'm sharing a different screen uh sets here but they are virtually identical you can use any so it's just
that they are given in two different formats as CSV and Excel but the data inside is the same. So let's take a look at one of those. Hopefully you can see the screen with the the data set. So you can see there
are there is even bigger data set here. There is um 100,000 plus almost 101,000 There is um 100,000 plus almost 101,000 um data attributes data uh points here.
So for each there is an encounter ID which is the uh specific uh admission ID. You have a patient identification. Uh the race of the patient is given. Uh
Uh the race of the patient is given. Uh the gender of the patient is available. The age of the patient in terms of brackets. So it's 0 to 10, 10 to 20, 20 to 30, 30 to 40, 40 to 50, so on and so forth up until you can see 90 to 100 I
forth up until you can see 90 to 100 I think. uh that is there weight is there think. uh that is there weight is there but you can see for but you can see for number of the
missing data elements. ments elements. ments uh admission type ID is there discharge
uh admission type ID is there discharge disposition ID is there is there uh pair code is there again that's missing for a number of cases the specialtity pediatrics internal medicine so on and so forth number of lab procedures number
of procedures uh medications outpatient emergency um inpatient what are the different diagnosis. Um number of diagnosis uh
there are then test results things which are uh which they have checked um um and then um so there is a bunch of those number of these uh including insulin um
right the admitted information it can be no it can be uh less than 30 and it can be greater than 30. No means there was no readmission. Uh less than 30 meaning there is a readmission within 30 days from the first first discharge. Uh
greater than 30 there is a readmission um after 30 days from the first discharge. So that's the um the data set. Now let's take a look at the uh the problem statement. What are the things that we
statement. What are the things that we need to do?
files. Uh one will have more details. Let's go through the first one first. This is a highle project view. Uh we already went through the objective. So we want to view that again. Uh we also looked at this. This
is the same. The industry relevance remains the same. The tasks remain by and large the same. It is a context that changes. Uh the outputs will be different because this is a different context in a different domain.
The references are also provided here. So you can you can use this. So So you can you can use this. So um this is the um the first set of details. Now let's look at the detailed um steps for completing the project.
introduction. We don't need to go through. We already had that view and tasks that we need to do? The first is BRD which is the same as uh the the the
BRD which is the same as uh the the the credit card fraud uh project. Uh then there is Excel task. You will see they are similar in the sense that you need are similar in the sense that you need to summarize key metrics. So things like
uh you know number of lab procedures uh what are the medications and uh what are the medications and encounters etc. Uh readmission by gender and age group. Race distribution. How many people are what what race are
there? What is the frequency of um uh readmission for that? uh plot patient count for each category of readmitted which is less than 30 greater than 30.
Uh you can use AI.t generate summaries and for uh suggesting policy actions. Uh of course you can use chart GPT for pro prompts. Um the this basically concludes
the Excel tasks. Then again on SQL task you need to create a schema. It should be called healthcare. That should be the default schema. And then create the tables using the diabetic data CSV.
Okay. Use CSV file to create a ski uh the the table. Then there is a bunch of things that you need to do within the SQL for data exploration. Calculate the
total number of patient encounters, top 10 diagnosis, um average length of hospital stay for each type of admission, number of readmitted patients and percentage of
total encounters. Identify the age distribution of patients, most common procedures, average number of medications, average number of medications, distribution of uh readmission across
distribution of uh readmission across different pair codes. Though those are the activities that you need to do for your SQL tasks. And need to do for your SQL tasks. And lastly for PowerBI tasks uh again you
need to create a uh distribution of readmitted counts of diagnosis and emergency cases through a bar graph. They have given how to do it. Visual They have given how to do it. Visual capturing the age bucket count by the
readmitted. Uh visualize the number of medications procedures etc. Uh create a chart showing the number of medications lab procedious diagnosis and use reubmitted as a color code.
Use quick measure suggestions to autoc calculated the average length of stay per diagnosis. Visualize the autogenerated DAX measure in a column chart. These are the tasks that you have to complete from a um from
a PowerBI perspective. Okay, the submission is the same. You create a single document with all these uh outputs uh kind of pasted into them
and then you submit in all three uh documents that are requested for your submission. Uh let me share my screen and I will go through the
and I will go through the um the the retail analytics predicting customer responses and optimizing promotions. Uh so objective of the project is to uncover customer behavior patterns and
predict responses to marketing campaigns. uh again you have to create a BRD do use SQL, Excel, PowerBI etc for uh enabling
datadriven strategies for what is segmentation? So segmentation is when you categorize your customer into different segments, right? So uh there
can be it could be in terms of uh their spending power, it could be the kind of family that they have. Um it could be any of those categories. So segmentation of customers campa camp campaign
of customers campa camp campaign optimization which is uh you target specific campaigns for specific customers. So that's another um objective that you have. Now you have promoscope which is a retail analytics
promoscope which is a retail analytics partner. Uh they are focused on customer behavior modeling so that you can optimize the campaigns. uh there are retail businesses which needs to uncover patterns in the
interactions uh so that they can make decisions right so for example um let's say you are um so let's say
um let's say you are um so let's say a family of four people where uh the parents are in their mid-4s and the kids are like teenagers what products do you think they will be interested in will differ from a family of two uh senior
citizens or somebody who is 60 70 years of age. Their preferences or the the the campaigns that they will respond to will vary quite a bit from the campaigns that
um the the other the first couple will respond to or the first family will respond to. So your ability to tailor the campaigns uh to your customers is often dependent on how well you do that data analysis.
Uh in a highly competitive market, understanding how customers responds to marketing campaigns is critical. Um business analysts use SQL, Excel and PowerBI along with AI power tools to transform complex retail data into
actionable insights which will help these different um uh different uh objectives. Uh your task is to analyze marketing campaign data to evaluate customer behavior, segment response patterns and
generate predictive insights to optimize campaign performance. Before beginning campaign performance. Before beginning the analysis, you need to create a BRD um which will be the foundation for the analysis. You will use SQL to do uh to
extract, join and structure customer and campaign data. Excel for exploring campaign data. Excel for exploring patterns um and PowerBI to build interactive dashboards. Okay, you can get the complete problem
statement and the data set. So let's take a look at the data set.
>> [snorts] >> Okay. So, we have a data set here which has a a total 2,200 a total 2,200 um uh what do you call um uh data data
data elements. uh sorry not data elements there are two 200 two two 2,000 plus uh 2,200 um data points that you have uh there is ID year of birth level of education marital status income uh kids at home
teen teenage kids at home uh the date when they became a customer uh recency when they became a customer uh recency uh different products that they purchase uh different products that they purchase um number of stores they visit
um number of stores they visit web visits that they do um catalog purchases, web purchases, deals purchases, all those uh things are purchases, all those uh things are there. Uh then you have um
whether they have accepted certain campaigns, the number of campaigns that campaigns, the number of campaigns that were done which is uh going to be which were whether those were accepted. Was there any complaint? Uh
there any complaint? Uh revenue from the customer uh response and the age. So these are the attributes that you have for this particular data set. Okay. Now let's look at the problem
Okay. Now let's look at the problem statements.
Okay. Um again same objectives same description uh details of the scenario description uh details of the scenario uh industry re relevance uh tasks that same. The output screens will vary because it's a different um uh output
because it's a different um uh output altogether. Uh the references are also given in this file. Let's look at the detailed uh problems or detailed tasks which will also be quite similar but we will go through those as well.
statement we already saw the first is BRB again the same template same information uh but with this context so you need to complete that that's the you need to complete that that's the first task uh then you have Excel tasks
uh create a summary of for the numerical features uh line chart for the number of features uh line chart for the number of enrollments by year uh crosstabulated count of responses value against education so you We need to look at
whether or not what what is the level of education and how many responses came education and how many responses came from that box plot on income histogram. Uh visualize response against marital
status um AI table to summarize a gum and create a prompt to find out which education group has the highest response rate. So that's the Excel task for exploration
So that's the Excel task for exploration and analysis. For SQL, first is loading. Create a schema. Retail data. Uh make it the default schema. Create tables. Um the default schema. Create tables. Um and then uh load the customer uh file uh
and then uh load the customer uh file uh as the data and create a um uh create a table and then do a number of calculations. uh and that will be the uh expectation from the data prep-processing for SQL
from the data prep-processing for SQL tasks. Finally for dashboards uh income distribution, year of registration. So that should be the first dashboard. Uh education levels and marital status that should be the second one. [snorts]
Relationships between income and wine spending that's your next dashboard. Quick measure suggestions to calculate average wine purchases per education group. Auto generated DAX measure in a column
chart. So these are the outputs that you need to create for uh the PowerBI tasks need to create for uh the PowerBI tasks and then in the uh uh word document you
can submit those uh so that you can complete the project. learners let's say AJ Abini Aira Alisha why don't you switch on your videos because you know when we begin with the session we discuss together it's always good to
look at each other and then talk all right so today's session it is not at all a PPT based it's not at all related anything to theory we'll only be discussing live examples from our experience as working professionals all
right therefore it makes a lot of sense to look at each other and then talk all right so we have got seller partner from Chennai, Hyderabad, Bank of America. All right. Kagi, you shifted from an from the services sector to the client side.
Maybe from Infosys or Accenture or you joined Bank of America directly. right, I see learners. We've got a good diversified pool. Okay, directly. We've
got a good diversified pool of learners and we have joining from all over the country. I see many of you joining. Thank you Shika for switching on your video. I see lots of people joining from south. There are some people who are
south. There are some people who are from north as well. Great. Great. All right. Sure. Learners, you know, I have a very simple quick question for you all. Right. If you could just share uh see my screen. I'm switching on my
whiteboard and a very small deck over here.
>> Yes. >> Has anyone heard of any of these three terms? Let's say you know if I ask you as a layman, people from bank also would also be well aware these days. So anyone who has heard okay Supa said no s I've
heard of these terms and maybe you would like to share one or two lines about what you know about that's it. Business automation is basically the way you automate the business of the company. The operation businesshu
The operation businesshu okay good sesh we've got someone joining from all state also. Okay P great what else learners can anyone else thank you else learners can anyone else thank you Sudesh anyone Santo is there as part
yeah so I have worked on a tool called as blue prism so which is basically an RPA tool >> RPA tool I I was actually just involved >> RPA tool I I was actually just involved in making the process documents PDFs
>> okay yeah so anywhere by any chance >> uh not automation anywhere but uh otherwise like business process uh like BPSM tools like APN I have worked with
so I was a developer few years back and right now uh in my current organizations like we are trying to build an in-house tool uh which is which works very similar to appen like where you can like
workflows >> got it got it something similar to Zapier okay thank you Nikl Uh next next anyone know why don't you pitch in if you've heard of even process mining is fine
maybe someone from bank we got two three people from the banking sector right they would be knowing much better these days about RPA or business automation or process mining anyone would like to pitch in
with very simple problem statement over here all I'll give you a minute. Read this and then I'll get back. I'll give you a minute to read. All right. Once you have read, why don't you just drop a yes in
the chat box. All right. So, just take a minute. It says you're a BA. You're working at an e-commerce. Uh, you know, your team spends hours compiling your team spends hours compiling customer feedback.
Great. Now learners what we'll do is I'll keep the problem statement open for you all to see. All right. And I'll give you another one minute. What you need to do is very simple layman terms. Maybe you know you do it on a daily basis in
your sector. For example, as an R as RM, you would be taking lots and lots of customer feedback. Sometimes related, let's say the app isn't working, this feature is not working. Sometimes people would come in and say something related
my permission, so on and so forth. Lots of lots of customer feedbacks, queries coming in day in and day out and you may be handling it in different ways. All right? So don't worry, you're aware about automation, process mining,
working professional, you're getting this problem. I'll give you a minute. Think of a solution. All right? Whatever solution you can think of. Maybe you can say that hey Sri why do we you know uh take data from all of these sources
in which I press a button in the morning. It automatically collects data from all of these sources, segregates it, analyzes it and gives
gives me a very beautiful report wherein I can just see the outcome. Could be an learners. Think of what all answers you can think for this problem. Maybe you can think of an example as well while you're explaining. But you got a minute
learners. All the best. I'll just keep quiet for quiet for a moment.
thinking learners. But as of now, what we'll do is we'll just try to listen to some of the ideas that you have got over here. Okay. So, let me know if you can see my whiteboard. And here we go. Let's start. Learners, go ahead. How do you
think you want to take this problem up? Anyone? Go ahead. Let's start. >> Yeah, probably like all these data this could be quite subjective. So maybe like
quantify these data >> uh or classify this data and then try to come up with some sort of a trend analysis or uh um maybe uh deviations or
>> Got it. Got it. So you set some parameters. List Nik is saying you set some parameters. Let's say on the ratings part or on something related to analyzing what this data really means and if you see too many negative words
horrible poor service so on and so forth so forth that can be highlighted so that this is something you need to focus on hinting at correct >> uh yes and like maybe like the kind of
problems that we are getting or kind of complaints we are getting so maybe >> classify I >> yeah classify quantify uh maybe like uh
maybe we can try and like assign some weightage to the keywords >> five okay so let's say if more and more complaints we are getting or more and more answers we are getting >> related to customer service all right so
that those complaints can be classified as customer service which shows that this is the area we need to work on more all right good nil next learner Next >> hello >> yes over to you
case if you're conducting a survey there need to be a base right for what you're >> you need to have a proper filters for example if it's running a customer statement >> uh if the problem like is it a customer
dependency or is it a system dependency right uh if it is a for example if it's a customer dependency how we can solve it it will be a different al together >> if it is that we can handle it internally. So it will be classified. So
I think uh what the first person said it it will be the same. So we need to put a proper filter to understand the journey. >> Got it. >> Good. So now another similar option it says that you know to work out the
feedback or details much better let's have a concept called filters and right Tina and once you have a concept called filters then you can >> even a good point on this Tina as you mentioned you can put who's in charge if
it's a system that's >> sorry in power or in Tableau we get an option you know where you can see the trend you >> yes Yes. >> So we can do that also if we can get the
analysis properly. >> Perfect. So let's say filters you came across a very big problem as you know Nitesh also mentioned it's customer service. All right. Now you can hand over this problem to the right team
because as a BA you're working across teams right so now instead of you being the person taking care of all the problems now you know that this problem has to be handled by Tina this problem has to be handled by Nickel's team this
problem has to be handled by Sevitast and this way you are able to narrow down and put accountability good answers Tina and Nesh anyone else idea over Hi sir, good evening. Kavia here.
>> Yes, Kavia. Over to you. Uh >> sir, actually based on this definitely queries. Yes. >> Yes. >> So uh in that definitely I could imagine we will get multiple questions on the
we will get multiple questions on the same uh problem so that we can duplicate that. Uh >> excellent. So we can make them as a unique those unique uh problems should be compiled and then
we can share that with the relevant uh concern team or a person. another good point Kav is saying that Sri whenever we are doing this kind of Sri whenever we are doing this kind of an activity we see a lot of duplicates
right and these duplicates tend to waste a lot of time. Therefore, apart from classification and filters, we also identify these duplicates and we remove them so that instead of a mountain to scale, we know what all areas we have to
work on narrowly. So, good answer, Kavia. We'll take one more answer. Kavia. We'll take one more answer. Anyone would like to pitch in. Savvita, maybe you'd like to go. Snigh, you would like to go. Anyone
>> Hi this is Savvita here. Go ahead. Go ahead Savvita. [clears throat] >> Yeah I was working for these automation things like um wherever we are finding things like um wherever we are finding the um any um issues from the customer
right. So we used to get the requirements and based on the requirements we were like I was working for the clicksense visualization tool where um wherever we have this manual things and wherever we have this issues
right so that one >> we were uh improvising that in the means of trends >> showing the trend and the uh showing the
sales also and everything whatever we have the issues. >> Got it. Got it. Learners, thank you. Thank you so much. Uh, Savva learners, example where this [snorts] thing was tried and in one scenario it backfired.
Okay. How many of you have traveled recently or maybe you've you know or traveled or anything with Indigo? Just type yes in the chat box. Indigo, Indigo, Indigo customers. Indigo. Anyone?
Go ahead. You can just mention in the chat box. Yes. No. Now this thing happened. Thank you Vishal. Now this thing happened. A customer didn't have a good experience while he was traveling. Okay. He puts up a tweet tagging Indigo
and he says, let's say that customer over here is, let's say that customer is Tina. All right, Tina. So, you went and you had a post on Twitter saying, "Thank you so much, Indigo. I had the most amazing
horrible experience working uh traveling." And uh she put a very nice paragraph sarcastically telling how bad the sarcastically telling how bad the experience was. All right.
Now, now the bot interpreted words like amazing, best, unimaginable. All of these tended to be entering in a very positive area. [snorts] And the bot
commented, "Thank you so much for your feedback. This level of encouragement and support motivates us to work better so on and so forth. So learners yes these kind of things do
help you know uh save a lot of effort but many a times these things tend to not be as evolved as a human can be. Okay now 30 seconds you've got 30
seconds go to three go through these objectives and then I'll come back.
Once you have read these objectives, just type yes in the chat box.
We'll be covering as many as real world examples for all of the discussions that we'll have today. Okay. So if you have read this learners just quickly type yes in the chat box. Hurry up.
uh I'll give you another 30 seconds. Read this then we'll discuss together. Okay. It says use of tech to automate repetitive rule-based task. Hurry up. I'll give you 30 seconds. Introduction to business automation. Now everyone
listen to this very simple example. All right. We've got let's say Snhel in this in this session. Snigh are you there? this session. Snigh are you there? Snail.
>> Snail, let's say you run a shop. Okay. Snail, >> any suggestions? What shop would you run? Any any suggestion? Maybe a book store, jewelry store, anything. >> It will be a food shop.
>> Say it again. >> Fruits. Organic fruits. >> Fruits. Okay. Now learners, we have Snill who has her own
let's say every day she has some customers coming in, invoices getting generated. Um, and let's say she's also have because she's doing well, she has also taken a GST number. Every day there are multiple invoices being graced.
That's how usually a shop works. Okay. [snorts] At the end of the day or maybe [snorts] At the end of the day or maybe at the end of every week day let's say month
snail would like to know top selling products total revenue that she has got. Maybe she would like to know date wise revenue. Maybe she would like to know let's say you know uh which day I had
highest sales so on and so forth. Do you agree that these kind of questions are very common for any business owner? Yes or no? Very common. Yes or no? Go ahead.
She has a name over here. Let's say Snighel Enterprises. Over here, she has a GST number. She has a date. And over here, she mentions the name of the product. Over here is the amount.
Over here is the total that the user has to pay. Let's say she would like to know how much sales were done on every day at the end of the
done on every day at the end of the month. Now a manual job would be open these invoices, copy this total, copy this date, put it in an Excel
this date, put it in an Excel and then it gives you a table. would work for you. Yes or no? >> Right >> now, if she's running a shop, she would be having almost the same format of a of
an invoice. Yes or no? It is not that she will change the format of her bill every day. Maybe once in a year or so she'll do, but she'll keep the same format of the bill. A very simple rule. Copy total,
A very simple rule. Copy total, copy date, paste it in a table and then give me the grand total. Do you think it is a very simple, straightforward logic? Do you agree? Yes or no? Does not require rocket science.
I mean even a schoolgoing student can do that. Do you agree? Yes or no? D, you find it confusing. Okay, maybe a high school student can do that. do you agree that this is a very straightforward logic of copying date
copying total and putting it in a format and putting in a table kind of a thing in this scenario learners since it's a very simple straightforward rule based we don't need snighel to waste her time in copy pasting
a very simple software can identify this is the date this is the total amount account and automatically create a table for her. Today, this same thing can also be done using chat GPD.
However, when chat GPD was not there, there were many other tools who can get this thing done for you. They can open your PDF, they can select whatever you want them to select. Let's say you have identified date, they can select second
thing as total and then they can keep on adding, adding and adding for you. Do you see that we have saved a lot of time and effort for snail? Yes or no?
something. Let's say, but do you agree that we saved a lot of time and effort for snail? Just type yes in the chat box.
My next question for you is my next question for you is please take let's say 45 seconds to 60 seconds. What you need to do is jot down multiple other benefits that snail will have when
other benefits that snail will have when she decides to use a software to get this thing done rather than manually copy pasting data every end of the day. I'll give you let's say 45 seconds more learners go ahead think of multiple
benefits. Maybe you can say that you know now Sneighhel is free from her time of doing this thing. She can invest that time in expanding her business. Maybe you feel that better insights would help her buying the right products. It's up
to you because now the software is identifying your top products. It is identifying your least popular products. It is identifying let's say which are your loyal customers because your invoices also have contact numbers. So
now she's getting tons and tons of data being analyzed and presented in a beautiful report format for her. And over here learners let's check to example that snail had told snail you had told right organic fruits if I'm not
>> that's right >> organic fruits all right do you let us remember one thing it's a perishable product right snail so you can think
something on supply chain as well that you know she can order her fruits faster so that they not she does not need a coal storage just to ensure that they do not get what you say spoiled. 30 seconds more learners. After that we'll discuss
some of the benefits that Snigh would achieve if she starts using this. business improvement efficiency uh uh accuracy saving time and efforts.
I'll give 10 more seconds learners. Then you need to give me some quantifiable and specific example related. Let's say snhel you have mentioned accuracy less investment on people work that's right great now if it is giving you some
accurate answers how is it helping you in what sense let's say sury it helps me understand that what time of the year what fruits are being sold maximum so that I can place orders in advance this is one example of benefiting from
accurate data learners when we talked about snhel's shop we had a Pretty straightforward. What do you say? Standardized process. Don't you agree Standardized process. Don't you agree that this is the invoice.
This is the product name. This is the total amount or let's say over here this was the discount that you know she offered. This
is the date and it will remain like this. The and it will remain like this. The software has to fetch only these details software has to fetch only these details and then the software has to put them to
understand what trend are we looking at maybe some questions let's say top selling products worselling products best sales so on and so forth which are
something that snal will use month on month it will not be a scenario that this month she wants to know okay what is what product is my top selling like to know this. It's something she would like to know on every month on
month-on-month basis. Therefore, do you agree learners that there is a very straightforward standardized process defined that our friend which is let's say robot we call our software as robot let's say has to follow. Do you agree
learners? Yes or no? Just type yes or no that yes is straightforward process. You copy paste this data analyze and then show it to me in the form of a trend. I tell me your names and you know where do you work what what role do you work at
where are you joining from do you think this is also a very standardized process yes or no very yes you're right Vishal it's kind of a standard operating procedure right therefore when you have these kind of standardizations
these kind of standardizations it becomes your first step it becomes your first step towards business automation although we have about machines today that keep on learning learning learning.
I know there are models that keep on learning and then they keep on adapting but we'll not go to that level in this session. Okay, because that would take another round of discussion. Now we here with the first step which is
some kind of a standardization is needed. All right. Now learners consider this scenario. Today our friend Jatin comes to Snighel and Jatin says hey Snhel you know you know we had a discussion about your software in
session snail says yeah so Jatin says you know what give me 1 cr I'll make this software for you do you think snighel will agree jatin do agree everyone that snel would
ahead and do that you know let's let me give you 1 cr and you give it to me tomorrow so that I can start using this software on from a food shop. The chances are very less that immediately she'll get convinced that you know this
will actually work. Is it even worth the money? So on and so forth. She's very However, Jatin says why are you being skeptical? Yours is a rule-based simple job. There's no questions asked. There is no brainy thing needed as such. Why
is no brainy thing needed as such. Why are you worried? Tell me learners how should Jatin convince Snhel that this thing is actually useful for her. Let's say Snighel is again from a non- tech non-software background. Jatan
would like to convince Snhel that this looks good. Learners how can he do that go ahead quantify benefits. That's great you know show her that you know these all are the benefits that you will be getting if you deploy this software.
Great. What else? What else learners? What else I think can do
you know Snehal agrees okay I think uh if we don't worry about the price also if we don't worry about the price also yes this might work let's say 1 cr for snighneighhel is not a big amount she runs reliance
hypermart for her 1 cr software won't make much of a difference but she's has to be convinced that you know this thing will work you think it will make sense If Jatin can give
a demo and one week of small trial so that Snehiel can understand what benefits will she exactly get from her
own system. Do you agree? Yes or no? Learners prototype. That's right. by you know giving her access he prototype app by you know Jatin can say you know you take this app try it for one week or maybe a month if you feel it's good then
maybe a month if you feel it's good then come back to me otherwise it's fine therefore learners before we ask snighel to deploy this concept at scale to deploy this concept at scale we need to test it
in a very small environment maybe if she's running something like reliance hypermark then maybe she can run it for one store one store works good she can go for 10 100 and
that's how she can scale therefore learners this is what we were discussing over here that you need some kind of a process standardization you need some process standardization you need some kind of a before you reach this scalable
part from process to scalable you need some kind of alignment that this is maybe your sales team is able to understand your supply chain team is able to make out some sense from this insights and most importantly they use
it because let's face it it's very difficult to change habits if someone is used to excels someone is used to doing change it they'll say oh wow this looks so beautiful but I will come back to my
excel therefore you need some kind of an alignment within your team so that they accept it and of course you formalized these requirements that I want these kind of insights. Top products, worst
products, highest revenue, date wise revenue, trend analysis, so on and so forth. Now all sorry now all of this that you see again works around whether the prototype was
successful or not. That means when you ran a Jatin ran a prototype with snail's big let's say Reliance Hypermart or fruit shop did she get the required
evaluation successful. So that's how it works learners instead So that's how it works learners instead of deploying it to scale you undergo a journey and this journey timeline varies across organizations.
Reliance, it may take maybe six months for a complete end to-end deployment. But if you talk a very simple uh let's say store, let's say a nearby hypermarket, then this thing can be implemented maybe in one quarter itself.
Again, uh very simple differences. I'll give you maybe another 20 30 seconds because this is just acting as a refresher to all the concepts that we are discussing through our examples. All right. Okay. Okay,
don't need to memorize or something. All it says is what we need to do understand what we are trying to do over here. Let's say we are trying to do some
kind. Don't worry about learners of these big big jarens process mining. Don't worry about these things mentioned here. All it says is if Jatan has to convince Snell because Snihel will pay a very good
amount to get this software deployed. Everything has to be documented. Do you Just tell me. Give me a quick heads up. Maybe you can give a thumbs up using a reaction. Maybe you can type yes. Maybe you can speak. Yes. Har. Do you think
that everything has to be documented so that after the deployment of work is done can't say you didn't do this right you agree that everything has to right you agree that everything has to get documented this is where
we see these steps coming in the picture first thing obviously would be Jatin first thing obviously would be Jatin telling Snehil that this is all I can deliver because I've gone through your questions I've gone through your
understanding and what exactly you're looking at so So this is what I can looking at so So this is what I can deliver. Good. Now before we go ahead sign document and you know go with the payment terms and all let's document it
in the form of a very simple document. Okay. This document Okay. This document will be followed and validated. See this is what I told you I will give and this is what I will give in the form
of a software. Okay. This is what it means over here. First Jatin tells documents and then starts building so that this particular thing learners developing a tech product or a software
has got infinite scope. Do you agree learners that today if I ask you to make a software for let's say Reliance Hypermart or for lifestyle or for H&M or for Marks and Spencer's home center you will have infinite number of ideas. Do
you agree yes or no? Just tell me yes or no. If you know we can do this we can get more loyalty points. We can have collaborations with let's say quick commerce. We can so many things could be possible when you talk of developing
some kind of technical softares. Yes or no learners. All right. Therefore at some point you need to limit your scope and this is what this business requirement document does for
you. It limits your scope that only this thing has to be delivered. Yes or no? Just quickly type yes in in Yes or no? Just quickly type yes in in the chat box if you have read this.
you have recently been so we've got people who have been to decalon right learners. Now listen to this. Let's say
learners this is what you would have definitely seen. There are two types of anyone tell what are those two types of checkout counters? What Decathlon
thought when it understood that you know let's create something like of a selfch checkout. So let's say Decathlon said all right you know let's understand if checkout counter what exactly she does. Step number one she'll come let's say
she's there the customer arrives. Step number two customer puts up his or her basket. All right. Now once she sorry yeah once the customer has put up a basket then
the customer has put up a basket then Issha would pick out each item scan it and put it away. and put it away. The software will keep on
The software will keep on identifying details cost this is the total bill this is the item code so on and so forth and finally [snorts] she'll ask isa isa will ask do you want
to have a carry bag or not very common question and then final bill is generated do you agree that roughly this is the process that is followed at any checkout out counter. We have taken Decathlon as
>> Yes. >> Right now Decathlon realized that we've got so many steps over here. Let us see if something can be done in the automation part so that this entire thing can be done using a selfch
Therefore what they said okay the customer arrives great that's step number one. [snorts] Step number two, let customer put items on his or her own in the basket. Step number three, the
in the basket. Step number three, the basket should analyze using an RFID basket should analyze using an RFID scanner or a tag. This is the item. This is the details of the item. This is the price. Add it in the what do you say
[snorts] screen so that the user can see this item has been added. Step number four, the user keeps on adding items. Step number five, there are some safety checks, right? That hey, you know, uh the user should not
let's say add an item and just take it out. So let's say some kind of a safety checks are present and finally the billing statement and all is done. Therefore learners when people were developing this product they analyzed
developing this product they analyzed these steps which is a part of your checkout process and that's when they created a cell
checkout process. Therefore learners this process that you know I would like to understand first what is being done manually how it is being done manually then only I'll be able to create an automation software or something for
automation software or something for that it is called process mining being done now in simple words we call it asis
it asis and where do you exactly want to be and where do you exactly want to be that can be your 2B over here. All right, this is very simple example of how process mining works. Now this
of how process mining works. Now this same thing can be applied to any process where you feel some kind of an automation could be a possibility. Now learners if you are able to understand what is process mining just
now if you're clear that yes Ryan now I understand the meaning of this jargon hurry up just drop me a yes in the chat box so I can go to the next one because I want you to understand where these things are applied and uh that's
when you'll be able to understand better okay
the end of this course we hope this session helped you understand business automation, process mining, RPA, BRDS, AI tools and how intelligent automation can reduce manual work and improve business decisions. But remember guys,
It starts with understanding the process, finding the repeated steps, identifying the problems and then building a solution that actually helps the business. If you have any doubts, question, ask them in the comment
forget to like this video, share it with someone who wants to learn AI and automation. Thank you guys for watching this video.