---
title: 'Agentic AI in Cybersecurity: Strengthening Defenses with AI Agents'
source: 'https://youtube.com/watch?v=EYG1hng4qhw'
video_id: 'EYG1hng4qhw'
date: 2026-08-08
duration_sec: 3819
---

# Agentic AI in Cybersecurity: Strengthening Defenses with AI Agents

> Source: [Agentic AI in Cybersecurity: Strengthening Defenses with AI Agents](https://youtube.com/watch?v=EYG1hng4qhw)

## Summary

This webinar, hosted by Simplilearn, explores the impact of agentic AI on cybersecurity, featuring insights from Harsharan Panuga, a site reliability engineer at Morgan Stanley. The session covers how AI agents can autonomously detect and respond to threats, the differences between traditional AI and agentic AI, and the evolving roles and skills required for cybersecurity professionals. It also highlights the growing investment in AI-driven security operations and provides practical advice for career advancement in this field.

### Key Points

- **Introduction to Webinar** [00:26] — The webinar on agentic AI in cybersecurity is introduced, with a host welcoming participants from various countries.
- **Host Introduction and Ground Rules** [01:38] — The host introduces herself and outlines ground rules: questions should be posted in the Q&A box, and certificates will be provided via a poll at the end.
- **Audience Engagement and Simplilearn Overview** [03:14] — The host engages with the audience, asking about prior webinar attendance, and provides an overview of Simplilearn's mission and achievements, including 15 million learners and partnerships with institutions like Michigan and Virginia Tech.
- **Speaker Introduction** [06:07] — Harsharan Panuga, a site reliability engineer at Morgan Stanley, is introduced. His expertise includes cloud infrastructure, DevOps, automation, and AI-driven engineering.
- **Scenario: SOC Alert at 2 AM** [10:46] — The speaker presents a scenario where an AI agent detects and fixes a security alert at 2 AM without human intervention, illustrating the potential of agentic AI in automating SOC operations.
- **AI and Human Collaboration** [12:58] — The speaker emphasizes that AI will not replace humans but will replace those who do not understand or leverage AI. AI requires human intervention in complex situations but can act autonomously when built correctly.
- **Why AI is a Top Priority in Cybersecurity** [14:08] — Attackers are using AI to accelerate attacks, reducing attack time from hours to seconds. Defenders must match this speed, making agentic AI essential for rapid defense.
- **Traditional AI vs. Agentic AI** [15:05] — Traditional AI (e.g., ChatGPT) waits for user prompts and provides suggestions, while agentic AI pursues defined objectives, acts with limited supervision, retains context, and completes tasks in multiple steps using tools.
- **Agentic AI as a Top Threat** [18:30] — 48% of security professionals identify agentic AI and autonomous systems as a top emerging threat by the end of 2026, highlighting its dual nature in both attack and defense.
- **Autonomy and Identity Challenges** [19:31] — Agentic AI introduces non-human identities requiring dedicated access controls. The more independently AI acts, the more securely it must be governed.
- **Regular AI Assistant Characteristics** [20:56] — Regular AI assistants wait for user prompts, answer one request at a time, suggest actions, and depend on the user to act. They operate within a single interaction.
- **Agentic AI Characteristics** [22:25] — Agentic AI pursues defined objectives, completes tasks in multiple steps, uses different tools, acts with limited supervision, and retains context across the workflow.
- **Q&A: Penetration Testers and SOC L1 Roles** [26:08] — The speaker addresses questions about job security: penetration testers and SOC analysts will not be replaced but will be augmented by AI. AI handles repetitive tasks, but human judgment is still needed for complex decisions.
- **Why Organizations Invest in Agentic AI** [30:01] — Organizations invest because attackers use AI, defense must match automated attack speed, there is a talent gap, and AI skills are becoming required for tech roles. SOC centers need more talent or automation.
- **Attacker vs. Defender Approaches** [33:15] — Attackers use AI for automated reconnaissance, prompt injection, privilege escalation, shadow AI, and memory poisoning. Defenders use agentic AI for threat hunting, security monitoring, alert correlation, and incident response.
- **Adoption of Agentic AI** [38:48] — 67% of security professionals already run agentic AI in depth, according to a Darktrace report for 2026, indicating that those not using it are falling behind.
- **Career Impact of Agentic AI** [40:41] — Roles are shifting from manual to supervisory, from reacting to directing AI agents, and from single-skill to dual-skill. Employers are hiring for AI fundamentals, cloud security, security automation, threat detection, and identity management.
- **Market Growth and Career Opportunities** [43:11] — The global cybersecurity market is projected to reach $562.72 billion by 2032, with a 62% growth in cybersecurity job roles over the past five years. New roles like forward deployment engineer are emerging.
- **Simplilearn Program Introduction** [44:38] — The host introduces an AI-powered advanced executive program in cybersecurity, delivered in collaboration with IIIT Bangalore, highlighting features like AI mastery, job assistance, and industry partnerships.
- **Q&A: GRC Career and SOC L1 Automation** [53:43] — The speaker advises on GRC as a career for freshers, noting it's good for those who like tech but not coding. He also explains that L1 tasks are being automated, but AI struggles with judgment, so humans are still needed for complex cases.
- **Agentic AI in Production** [56:36] — The speaker shares that agentic AI helps in triage and diagnosis, reducing alert noise and resolution time by 30-50%, but full remediation still requires human approval.
- **Closing Advice: Consistency** [01:00:10] — The speaker advises building small atomic habits, showing up daily, and not missing two consecutive days, as consistency is key to learning and personal growth.

### Conclusion

Agentic AI is transforming cybersecurity by enabling faster, autonomous threat detection and response, but human oversight remains crucial. Professionals must upskill in AI fundamentals to stay relevant, and consistent learning is essential for career growth in this evolving field.

## Transcript

YouTube and LinkedIn. So for anyone joining us from YouTube and LinkedIn security webinar. The topic is agentic AI in cyber security. A very interesting topic to talk about. We also have a wonderful speaker joining us today. I'll
before that, let me take a look at all the introductions that have come in. Uh so there's Fred joining us from Tanzania. Hi Fred. Thank you for joining us. Hi Sophia. Hello. Hello Mihul from London. Rohit
Yadav from Mumbai. Hi Ragha. Hi Alex from Peru. Raja S from Hi Ragha. Hi Alex from Peru. Raja S from India. Hi Rohin from Pune. Hi Kumar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar ar architect with 18
years of experience. That's that's great to know. Hi Ammon. Hi Karthik. Hi Confidence from Nigeria. Great. Hi Pile. Hi Ravi. Hi Raindra. So many
wonderful introductions. Thank you all for being so active in the chat and telling us where you are joining us from. So it's a mix of audience joining us from across the world which is wonderful to see. So before we begin, I
would also like to introduce myself. I'm Schwatha and I will be your host for today's session. Uh we have a wonderful speaker like I mentioned uh earlier. So we have Hashwaran joining us. I'll introduce him shortly. But before that
uh let us quickly go through the ground rules for today's webinar. If you have any questions, I request all of you to post your questions in the Q&A box. So this will ensure that your questions are addressed. We have a dedicated Q&A
section towards the end of the webinar. So if you probably ask your questions in the chat, uh we might miss out on it because of the high volume of questions and messages we received there. So please post your questions in the Q&A
the end. Let's use the chat box effectively and let's refrain from sharing any external links in the chat. We would really like to keep these conversations related to today's topic uh also because the topic is really
interesting. And finally for the question which is always asked the most which is your certificate of participation we will be sharing a poll towards the end of the webinar. You will have to fill your full name in the poll
to receive your certificate the recording and the slide deck to your email ID and you will definitely receive that within the next 48 working hours. So these are our ground rules and I hope we are aligned on the same.
Uh you guys can give me a thumbs up or any kind of reaction. So that would be any kind of reaction. So that would be great from your side. that. Uh so let's move on. So before we go ahead, I just quickly want to ask all
of you, have you attended our webinars before? Have you been a participant or is this your first time attending a SimplyLearn webinar? If this is your first time, please comment or type uh first in the chat. If you've attended
webinars before, you can mention the number of times you've attended if you remember. Uh, okay. So many first timers. timers. Third time, first first. All right. Five
times. Second time, second time, first time, first time, multiple times. Great, great to see this. Uh, very, very happy to see a lot of you who've attended webinars before and attending again. That's great news for us. So you all
would probably know about simply learn but for anyone who's joining us for the first time once again thank you for joining us we hope you have a wonderful experience with us. I'll quickly take two minutes of your time to uh let you
know what we do at simply learn. So at simply learn our uh goal is to transform lives by empowering uh people through digital skills and we do these three uh through multiple courses that we offer in different domains that are there. We
have already helped over 15 million learners advance their careers across 150 plus countries and we do these through global partnerships. For anyone who has attended webinars before, you would already know the kind of programs
that we provide and our partnerships as well. So some of the key partners that we have are university partners and industry partners. some of the most reputed institutions and universities like Michigan Professional Education,
like Michigan Professional Education, Virginia Tech, Google, Microsoft, AWS and a lot more. And today we're going to be talking about triple ITB as well. Uh triple ITB offers a special course in cyber security about which we will
definitely talk about in the coming slides. Uh so there are some key metrics that we wanted to highlight here. something that talks about the kind of experience that you can expect from us in the next one hour. So, uh most of our
learners have reported a 50% salary hike on completion of courses with us. We have also been rated 4.8 out of five by most of our learners. So, this is the kind of experience we've delivered to all of our learners. And one of the um
main highlights we have is that we have a graduation rate of 80%. which is higher than most of our competitors also because uh we ensure that our learners stick with us till the very end. They don't drop off. We have a very good
support system that always helps them and we hope to maintain that as well. So these are some of the great highlights that we have and wanted to emphasize uh in this session. So before we move ahead, I'm sure all of you are waiting
ahead, I'm sure all of you are waiting to hear uh about our speaker and I am very happy um uh to introduce our speaker for today's session. So joining speaker for today's session. So joining us today is Harshuharan Panuga, a site
reliability engineer at Morgan Stanley with extensive experience in cloud infrastructure, DevOps, automation and building scalable systems. Throughout his career, he's worked across leading organizations, giving him a strong
hands-on experience in designing and securing modern technology environments. What makes Hara a great speaker for today's topic is his practical approach to AI and cyber security. His work
focuses on AIdriven engineering, cloudnative technologies and also applying automation to solve real world engineering and security challenges. So the insights he will be sharing today actually comes from real industry
experience and not just theory. So I'm sure that all of you are going to get a lot of value from today's session. So uh Bhasha we are really thrilled to have you with us. I'm sure everyone's waiting to hear more from you. So over to you.
>> Yeah thank you so much Sua. Such a great introduction. Thank you so much. Um and introduction. Thank you so much. Um and hello everyone. Good evening. Um everybody's have uh hope everybody is having a great evening ahead. So I
really appreciate all of you for taking time a valuable time apart from your work and you are spending your time to learn something or to get something um in your uh out of work hours that truly tremendous that is what I actually did.
So yeah I'm looking forward [clears throat] um for today's session. So as Shua mentioned today we are going to understand um how AI is actually affecting cyber security in the positive way and also in a negative way in
specific agentic AI. So how we can do uh autonom how can we uh apply the automations on cyber security because cyber security is a domain where AI it is very hard for AI uh to replace because it requires the human inter uh
human intervention the roles like uh uh infrastructure engineering and which closely are related to the um what I can say the infrastructure of the the systems. So it is very hard for the AI to replace but of course we can leage
to replace but of course we can leage with AI. So I wanted I want to listen the enthusiasm all of you on the screen everyone. So can you please show your enthusiasm with the chat section. So how excited are you today? So so that I can
give you the expectations for today's specific session. Meanwhile I will share my screen. Uh main value. Uh
Hope everybody is ready for the session. So I am also very excited. So So I am also very excited. So yeah.
So, Shua, it seems a lot of students are uh quite interested and of course, yeah, please acknowledge everyone whether you will be able to uh see my screen or not.
>> Yeah, your screen is visible. >> Oh, give me a second. um I'll walk you through the u how AI is going to revolutionize the era of cyber
security and also how we can use or how we can leage the agentic AI. Okay. So how we can uh leverage this um cyber security domain maybe it might be IT
security it might be anything with the help of agentic AI. Okay. So now I'm acknowledge me whether you will be able to see my screen or not.
acknowledge whether you will be able to see my screen or not? Yes, it's obviously important responses. >> Got it. Got it. So, yeah. So, now let's try to understand uh with a small example or with small scenario that
let's say you are in a midnight. So, you know, S right? Of course, a lot of people might be knowing about SOC security operations centers. So these security operations centers. So these are the centers where um if any uh
attacks or any cyber security incidents uh will be alerted to the specific centers. So the people in the SOC like SOC analysts or SOC uh people they will take care of those cyber security incidents and they will
try to figure them out. Okay. So let's say it's a night 2:00 a.m. you are in say it's a night 2:00 a.m. you are in India and a small alert or not a small a cyber security alert uh was fired in the SOC. Okay. So here let's let's imagine
that instead of waking you like you waking you at 2:00 a.m. in the midnight. Let's say we have an AI agent which spots the error or which spots the incident and threat and it recommends a fix within minutes and it also fixes the
fix within minutes and it also fixes the system. A machine is there for you which is helping you to automate your work. Not a machine maybe an agent. Okay. So revolutionize. It is or uh it is architect or
orchestrated such that any cyber security instant or maybe any SOC alert is going to is going to given as an input to our AI agent. What AI agent
will do? It is going to understand the whole alert. the description of the alert uh the cause of the alert everything it is going to it is going to read and then it is going to take a decision by itself okay harsha so some
security alert came so now what I need to do it is not going to ask any human it is going to take decision within itself and that to within minutes and it is going to fix okay such kind of operations that we can do using agentic
operations that we can do using agentic AI okay so now let's try to Understand this. The question isn't whether about AI will be a part of cyber security or thing that you need to understand everybody. Okay. So here you should not
think about whether AI will be a part of cyber security or not. It's whether you know how to work alongside it. You uh it's whether you know how to leverage AI to improve your productivity of your work or to improve the effectivity of
the work. So this is very simple. Okay. So here AI um never replace human in in in such complex situations. Okay. It requires human intervention. Okay. But
requires human intervention. Okay. But AI agents can be built in such a way that it is going to take the actions autonomously. Okay. So of course there would be some impact. It is not going to replace the humans. It is going to
replace the persons who are not able to understand AI or who are not able to follow AI alongside with the current domain. In our uh domain site reliability engineer the current domain I'm working in in the DevOps also we
will use AI to analysize to analyze the logs to detect the errors. Okay. So AI is everywhere but it is just that how you are optimally using it. Okay. So now let's try to understand the whole agenda for today's session. Okay. So now we'll
try to understand why AI is now a top priority in cyber security because attackers are now going to be tremendous. They are taking advantage of AI and they are fastened their attacking process. Previously it is take it is
going to take them a multiple hours maybe multiple days to attack a system. But now with the help of AI the attackers were tremendously able to attack that specific software within seconds or within minutes. So we should
also build system such a way that we should defend ourselves within seconds. Okay. That is why we are learning today about this agentic AI. Okay. So now we also understand previously we have traditional AI. Chart
previously we have traditional AI. Chart GPT normal is traditional AI. Okay. Whenever you write an email you just ask charg you give a prompt it will give you output that is called as traditional AI. So here
how different uh agentic AI is from traditional AI. So why we are not using traditional AI in our current systems? Why only agentic AI? Okay. Here we can discuss about the key terms like autonomy, decision making and goaldriven
execution. Okay. And also we'll we'll try to understand why different try to understand why different organizations like IBM, Google, different kinds of multi-ir multinational companies are investing
abundant amount of funds in AI uh AIdriven security operations. Okay. And also we'll try to understand where the specific security professionals fit as AI as AI is going to take more of the work because uh let's say before 2023 or
work because uh let's say before 2023 or 24 um the work of uh I uh what I can say SOC analyst might be a different thing but now it is more modernized more uh standardized and automated few things there are some change changes in the
roles and responsibilities Right? We also try to understand them. Okay. So now let's have a small question. So before we start, which describes best your role? Take the options over the screen. SOC analyst, IT
SM, IT security manager, Cisco related uh roles, security architect and maybe exploring or who are trying to switch to this domain and someone very curious
about this AI. Please respond over the chat section. Everyone
uh five and six means exploring and about very very curious about AI cloud not mentioned over here, you can write your role over there. It's it's your role over there. It's it's completely fine. IT security manager.
Cool. We have a lot of IT SMS and also about also we have SOC analysts and also about also we have SOC analysts and also we have a Cisco uh related roles. Okay. And also I'm very happy that a lot of professionals have a good amount of
interest on agentic AI and AI and they're exploring this stuff. I really appreciate all of you. Okay. So and I also appreciate all the the the individuals who are already in the roles of SOC analyst, IT security engineer or
manager and cloud architect because you guys are fast forwarding and you're also moving alongside with the revolution of AI era. Okay, because we should upskill ourself uh continuously because the
every day you can see a very new update in the era of AI. So that's why at least you need to make sure that you are you are um uh what I can say following with the recent stuff and also upskilling yourself as much as possible. Okay. So
now let let's move on with the first part. Let's in this first part we we'll try to understand why agentic AI is a biggest shift right now in across all the domains not only cyber security but also different domains. Okay. But we'll
discuss about uh about only security here like uh SOC related stuff cyber security. Okay. So if if I talk about the numbers this 48% of the top emerging
threat of security professionals identifying that agentic AI and autonomous systems are going to help the cyber security also adversely affect the cyber security in the year like at the end of 2026.
Okay. So here let's try to understand why agentic AI matters over here. Okay. So here attack and defense they are evolving together. So attack as I mentioned attackers are updating theirel with a lot they are taking advantage of
AI they are taking advantage of AI models and machine learning also in some cases and they are attacking like this they are attacking very fastly and defense also should evolve together. So that's why agentic AI came into picture.
that's why agentic AI came into picture. So here autonomy means um automatically detecting the frauds, automatically detecting the SOC alerts and mitigating it and trying to fix it. Here the autonomy plays a pivotal role in
detecting the issue and reporting the issue to the relevant stakeholders and taking the relevant action. Okay. And also here previously we don't have AI agents. So here we will be having only one identity check whether he's a human
one identity check whether he's a human or not. Now there is one more identity challenge. So here of course AI agents are nonhuman identities. Of course they need a dedicated access controls because we can't give all the access as same as
the person to the agents. Right? So agents we will be giving some specific part some particular access to them which they are supposed to work on. which they are supposed to work on. Okay. So here we we must understand that
the more independently AI can act the more securely it must and should be governed because if you're giving a free wings to AI you must take care of its
[clears throat] governance. It's it's a security uh operations because you must take care that how securely it is managing your tasks. Okay. you can give uh a full access to your system but make sure that access is secure.
Okay. So now let's try to understand what is agentic AI and what is regular AI assistant which is nothing but a AI assistant which is nothing but a traditional AI. Okay. So here let's talk
about the regular AI assistant first. Okay. So here the regular AI assistant Okay. So here the regular AI assistant it is going to wait for the user prompt. Of course, charg is not going to give you the answer until and unless you are
giving the input. Yes or no? Charge GPT is going to give you the in output if Right? Let's say what is the weather right now? Hey, give me answer for or rewrite this specific email or re rewrite this specific paragraph. If you
give input, then only charge GBT processes and gives us the output. Okay? processes and gives us the output. Okay? And also it answers for one request. One request, one answer. Okay. And it suggests what to do. Means hersa, you
can do this. If you if you are if you're asking to rewrite uh a mail, it is just giving you the uh draft of that mail. It is not going to actually send mail. It is going to help you. It is going to suggest you what kind of mail that you
suggest you what kind of mail that you can uh give with. Okay. And also it depends on the user to act whether you take the draft or not. Let's say charged or neatly corrected version of your email. Now it is on you whether you are
going to use that draft to use in your email or not. It completely depends on you and it operates within one interaction. Now let's shift our wings interaction. Now let's shift our wings to the agentic AI. Okay. So, agentic AI
pursues a defined objective. Let's say I am creating one agent. Okay, let's say I want to book a ticket in a railway. Of course, it is not done yet, but I'm just
giving you an example. Okay, uh let's say I want to book a railway ticket on the platform. So, here I'm going to build an agent which is specifically designed for booking Indian railway tickets or maybe any tickets, bus
tickets, train tickets. you will be giving one defined objective end goal for that specific agent. That is a main uh rule of AI agent where you must
uh rule of AI agent where you must define an objective or expected outcome define an objective or expected outcome from the agent. Okay. So what what is our objective in this case? It is going to understand the whole process and it
is going to book a railway ticket for you. It takes the input from you. Anything it takes all the things from you and autonomously it is going to book you and autonomously it is going to book a railway ticket for you. Okay. And here
is it going to complete the whole booking process in one single step? No. Right. First it needs to take the input from you and then it needs to search for from you and then it needs to search for the trains which are given from station
to station date and everything. So searching for the correct trains is one task. After searching for the after searching for the trains, next selecting one train and also giving all the relevant details, passenger details is
one task like uh automating uh like giving the capture and everything is one task and payment is another task. Okay, here you might be getting a doubt s how here I'm just giving you an example. Okay, it can't be done by AI like
automating the payment is not yet done because it's a risk of risk of monetary things. So that's why uh automation of payments is not yet revolutionized
because we must secure that. So the the teams are working on it. The AI teams are working on it. Okay. But currently the payments the payments are not yet fully automated. But now I'm giving you an example. Okay. multiple tasks will be
an example. Okay. multiple tasks will be there in your what I can say agent. Okay, it uses different different tools. It completes one specific task in multiple steps. Okay, it can act with limited supervision means you need not
to stay with it. Okay, you need not to stay with it. It is automatically going to work. Okay, and it retains the context. You know what is context? Context is nothing but it is a kind of a small memory where it will stores the
small memory where it will stores the user's chart history means chart context okay let's say if you go to chart GPT in one tab you are going to like let's say you want to build one small project you will give all the data all the details
about your project over there yes or no and in that specific chart it is going to understand uh everything about your project but if you go to create a new knowledge about that specific project that is called as context. So in agentic
AI so you can uh retain the context across the whole workflow. Okay. So now um let's try to understand why organizations are investing. Before I go ahead I just want to pause myself and let's see if any questions were there
from the students end. Okay open open for the questions. If you have any questions please do let me know or if everything I will go ahead. Uh Hashad, there's actually one question in the Q&A. Okay.
>> Uh so someone's asked, "So a penetration tester will potentially lose their job in the near future." >> No, not really. So here pentesters are see the cyber security domain is a domain where unpatchable by AI to
replace but not to increase the productivity. Okay, you can leverage UI. So I can tell you some domains like let's say a IML domain, cyber security let's say a IML domain, cyber security domain, DevOps and also S sur like AWS
or cloud infrastructure roles are yet to be touched by AI to be replaced. They are the they are the jobs that can't be replaced but you can make your job effective by using AI as simple as that. So pent testers are not going to be
replaced because you need to understand you need to analyze the upcoming attacks like the possible attacks that is the main that is one of the main jobs of the main that is one of the main jobs of the pentester right so to um what I can say
to um what I can say to uh what I can say to guess or to understand the impact like how attackers might attack it's like how attackers might attack it's kind of u you acting as a uh attacker
to check whether your website is vulnerable or not to checking your checking your system security by yourself not rather than attacks. So yourself not rather than attacks. So here no AI can uh replace the pentesters
but it can help you to make your job more easier. Okay. Uh so hope hope that makes sense. >> Yeah. I think as a part of that question >> Yeah. I think as a part of that question someone else has asked what about SOC L1
someone else has asked what about SOC L1 level? Yeah. So here uh here SOC alerts level? Yeah. So here uh here SOC alerts are the incidents are um let's say all the alerts are not identical. Let's say one alert is about something second
alert is something and every day you might experience the new SOC alerts. So might experience the new SOC alerts. So here you can build the AI agents and also you can leverage AI to identify the SOC incidents and SOC alerts and you can
take the relevant actions with the help of AI. But here AI cannot going to replace completely about completely SOC an okay as far as my experience and I the incidents they have seen across the multiple fonts. So here no uh
individuals are affected during uh with like no individuals are um uh affected by AI but they are benefited by AI because they are doing more jobs they they are leveraging and they were increase their SLAs they are going to um
increase their SLAs they are going to um meet SLAs very correctly like and also they're making sure that this type of things will not recur and they're taking very relevant actions. So it's kind of improved and it's kind of a what I can
say uh a kind of a very helpful teammate in the team. That's it. So it is here it is going to help you but it it should not replace your brain. It should be your companion. You should take the suggestions. That's it. You should not
replace AI with your brain completely. Okay. So nowadays I see some of my Okay. So nowadays I see some of my friends were um without charges, right? So previously we used to write we used to uh use Grammarly and different
kinds of platforms to rephrase our emails. Now now everything has changed. Now nobody's investing their time to just write emails because they they can use that time for different thing. So that is how here the time is going to
that is how here the time is going to optimize here. specific thing like why organizations are investing in this specific domain.
Okay. So here of course attackers are already using AI as I mentioned. So now defense must match the automated attack speed. If they attacking within seconds we should also defend ourselves within seconds and we should be capable of
seconds and we should be capable of doing that. Okay. And also we should be we should be making sure that we should stop the attacks. We should detect the attacks and we should stop them and we should mitigate them within seconds as
speed as the attackers. Okay. And also there is a lot of talent Okay. And also there is a lot of talent gap. Of course a lot of people know um networking. They know about very they are very good at Linux administration.
are very good at Linux administration. They're very good networking and um they tools. So they know everything about they have a certifications also SOC certifications but now we we should follow the uh what I can say the firm's
expectations and also the industry's expectations. So here it become like the expectations. So here it become like the AI skill maybe the basic AI skills are made required for any tech role nowadays. Now I see uh whenever like I
see some of the interviews that we were supposed to test the candidates's AI skills as well. Okay. The previously I was asked to check whether the candidate is able to at least learn AI because they should know basic stuff about AI.
So what is tokens? What is rag or maybe what is LLM? So basic stuff very basic. they should not be very expert in that but they should have at least the small uh like minimal knowledge expected knowledge. Okay. So and also whenever if
knowledge. Okay. So and also whenever if you whenever you see that um alerts are getting increased in these days. Okay. So now SOC centers needs days. Okay. So now SOC centers needs more talent or maybe if the talent is
not there of course they're going to um what I can say create automate their specific tasks. As I say for SOC the competition is comparatively low as with the software development I
say. So because of course that there should be there would be a competition between multiple domains but when it comes to the comparison between the SOC comes to the comparison between the SOC or cyber security domain along with the
what I can say uh the specific software development web development fullstack development the competition competition is comparatively low. Okay. So in like
is comparatively low. Okay. So in like and also the talent or maybe the source in the the talented individuals are are also very low because of course cyber security requires a lot of logical thinking, analytical thinking and and
thinking, analytical thinking and and also it requires that spontaneous that you should you should uh quickly understand the cyber security incidents and alerts and you need to make sure because every second matters over here.
Okay. So agents are going to help within seconds. That's why organizations are investing more in the agents or maybe in the era of AI. Okay. And this is proved in across the organizations like IBM and Google. Okay. So now let's try to
understand the second part. The two sides one is attacker and the second one sides one is attacker and the second one is defender. So now same technology here everybody is using same technology but the intentions are different. That's why
the new phrases attacker, defender, these things are coming. Okay. So here the objective or intentions are changing the outcome of that specific operation or specific instant. Let's say I also know Linux. I also know SQL SQL
injection and I also know I I know all the security operations. So now my intention is to attack the system, steal the data, sell the data, get the money. the data, sell the data, get the money. that is one object objective. So that is
making me as an attacker. But when it comes to defend defending uh here we need to make sure that using the same technology that no attacks should happen. Using the same technology our systems are secure.
Okay. So same technology with two different outcomes. Okay. So if you talk about the attackers approach how they are going to attack. So now attacks previously are going to take some hours or maybe some days or sometimes months
also. But now attackers are moving at a machine speed and they using they are using agents to scan adapt and act with minimal human input. So here they can minimal human input. So here they can use multiple agents to scan the systems.
Okay. So now if there are multiple attacks like automated rec recognizance. So here it is the the agent is going to map the attack surface at a machine speed machine speed. Okay. It is going to find the exposed assets ports
everything then teams can patch them. See to identify the security patches it takes multiple hours for the security analysts or maybe managers but agent is
going to resolve or going to conquer that specific task within seconds or within maybe minutes. Okay, prompt injection. So they are going to um use agentic AR maybe our AI tools. They feeds hidden instruction to a AI
tools. Let's say I'm using a very good prompt. They are going to inject very prompt. They are going to inject very wrong instructions to my AI and they are going to turn our organization's agents against it. Let's say I have given a
agent which is going to defend. I have given a very good prompt to it. Now what attackers will do they will they will inject very bad or very uh dangerous prompts to that specific prompt that we already given. Okay. So now privilege
escalation. Okay. So now it chains a smaller weakness to gain a deeper access. So here even one weak point in the system can lead to the wider system control. That's why systems like especially banking systems they are
going to take care of each and every system. Okay. So in our firm we are we system. Okay. So in our firm we are we are not using AI to write mails here. We are building agents. We have our firm's own AI. Okay. To make sure that
everything was onsite, not the external one. So here that's why firms are using to that they're hiring AI engineers to build their own AI agents or maybe chart
models for the own firm's internal needs. Okay. So now shadow AI means it runs unapproved AI agents without oversight. Attackers will build AI agent and they will run that AI agent in our systems
will run that AI agent in our systems without any uh uh approval and memory poisoning. It corrupts means um we are storing a context for our AI agent. So attackers will distract or will completely clean that context which
which leads to malfunctioning of our AI agents. Okay. So now let's try to understand the defenders approach means us how what are they going to do they are going to use agentic AI for the threat hunting to detect the threats and
to mitigate them and security monitoring they're going to monitor the systems continuously alert correlation so they connects the related events into one clear picture and cuts through the alert overload and they're going to take the
actions autonomously and we have the incident response which contains um uh which contains and isolates the isolates the threads within minutes. Let's say we have multiple alerts. You need to isolate you need to
segregate or you need to manage each and every alert based on its nature based on its um uh what I can say department or maybe based on its separate thing. Let's say the security is on backend systems or front-end systems or database or our
CACD pipelines whatever it is going to segregate. Okay. And also there are some SOC automation tools as well. Okay, it it automates the repetitive security workflows. Let's say you have some Linux commands. Uh, of course I think you will
be doing it using shell scripts. Now it is going to make and you are going to run that shell scripts using um chron jobs. You are going to automate them using chron jobs. But you can do more effectively using AI agents also because
chron jobs is a traditional agent. You can say something like that. But now agents are coming. So now teams are using agents to run the scripts or to using agents to run the scripts or to detect the things. Okay. So around 67%
of the security professionals already run agentic AI in depth. Okay. So if you're not of course you are already behind. The source for this number is a behind. The source for this number is a data. It's a dark trace uh which is um
displayed in the year 2026. Okay. So that's why it is very important to learn about the agents. Okay. So how agent is going to work? Okay. So now a agent is going to work? Okay. So now a small live poll. Everyone
I am I I'm displaying four different kinds of agents. Please tell me which of these agentric AI use cases was completely new to you. 1 2 3 four
in with cyber security AI right uh Satsua yes there are no restrictions for somebody to enter into cyber security career it's just that it's [clears throat] a matter of right skills that's it if you have relevant skills
your previous background doesn't matter a lot Okay.
companies who have developed autonomous pen testing using AI yet human touch is still needed to improvise it with the time do refinement etc. I completely uh support this uh harag statement because of course teams are using AI to automate
intervention to check whether it is correct or not. correct or not. Okay.
now what is this means for your career? Okay. So means u agentic AI what agentic AI means for your career. Okay. Let's try to understand that. So excuse me for
this misalignment. Okay. So here why the roles are changing. So here this agentic this agents are handling the volumes. So you own the judgment. Let's say it is
going to let's say thousand alerts came and you are assigned with maybe 100 and you are assigned with maybe 100 alerts. Okay. So now agentic AI scans each and every alert and suggests you to the fix. And here your task is to just
judge whether that um finalized decision by the agent is correct or not and just approving the AI agent and checking whether the what were the steps checking whether the what were the steps taken by the AI. Okay. So um from manual
to the supervisory you you were supervisor for the specific agent from reacting you are directing AI agents what to do from a single skill to a dual skill. Okay. So that is how it is going to actually uh impact your role. Okay.
So the skills employers are hiring for of course AI fundamentals, cloud security, security automation, threat detection, AM which is identity and defense fluency. After the cloud evolution like AWS, Azor
evolution like AWS, Azor um the security become much easier of course sometimes complex and also it opened a lot of opportunities because now anybody can take own infrastructure uh cloud infra infrastructure okay so
there's a basic skills that employers are looking for okay like AI fundamentals like how how how to build the agents how agents work what is the agents workflow, what is the agent's um
pipeline, agent AI pipeline, everything everything you you should you should be able to understand and it's a matter of your time. That's it. It is nothing no magic. It's no u rocket science. It's just you you need a a certain amount of
time every day that you will learn about these things. Okay. So now where this takes to uh where this takes your career. So here of course as I mentioned 562.72 billion projected global cyber security
billion projected global cyber security market by 2032 and 62% growth in cyber security job roles over the past 5 years. Okay, if you see the roles like SOC analyst, CSE means cloud security
engineers, security architects, cyber security consultants, threat hunter, new job. Okay. So forward now I see a new role is coming up within the firms like Google like forward deployment engineer like it's
kind of he's a he takes care of everything from sec from building apps to securing the apps. So the new roles are evolving with with AI. Okay. So now
what is next? So you know how AI agents are going to take care or how they're going to help us in future. So now what is your next step as a current professional or maybe the one who is willing to shift into this career or
maybe a student who is take who is willing to pursue this career path. willing to pursue this career path. Okay. So building the skills step by step it takes a lot of discipline and a lot of consistent consistent efforts in
your learning. Okay. So how you build these skills in a structured way. Okay. So now we have something a very beautiful program that Sua will explain you. I think I will give I will hand over this to Sua from here.
>> Yes. Uh thank you so much Harsha that was really insightful. Um I'll quickly was really insightful. Um I'll quickly share my screen as well.
visible to all of you. Okay. All right. Uh so like Harsha was me mentioning you all came here to understand the impact of AI in cyber
security. I hope all of you got great insights. Uh and we got a lot of questions about uh if whether people can actually get into cyber security without any uh without any background in this field. So with that we have uh a course
it's an AI powered advanced executor program in cyber security and this is delivered in collaboration with triple IT Bangalore. So I'll quickly walk you through some of the key program features that we have as a part of this program.
So if you're someone who's looking to build a long-term career in cyber security, it's very important to remember that learning individual tools is not enough anymore. So employers are looking for uh people who can actually
solve real uh problems and real life uh uh uh real uh problems by implementation and that is exactly what this program is trying to solve. Uh one of the biggest advantages that you don't just learn from cyber security uh but you will
learn that by using AI. So AI mastery is one of the major emphasis that we are giving as a part of this program. Uh secondly, we have AI powered job assistance as well. That is another key highlight that we are providing. So as a
part of this course, you're going to get access to AI powered rumés and LinkedIn updates. And also one is you want mock interviews because we all know that we've been in situations where you know the content but you don't perform in
interviews or something happens while you appear for the interview. So we have a system in place so you don't face any issues while giving your job interview. Uh we also have an industry partnered curriculum. So you will get C version 13
curriculum. So you will get C version 13 AI training and EC council exam voucher and also Microsoft learn access and badges as a part of this program. And finally like I emphasized earlier as well. So you will have access to 30 plus
global case studies and 100 plus guided assignments in cyber AI and 30 plus cuttingedge AI and cyber tools. You will also have a few master classes with the faculty of triple IIT Bangalore. So this is a great place to get the exposure all
of you are looking for especially in the field of uh cyber security. So if you are to look at um the you know core modules that are covered the curriculum is actually designed to take you from the fundamentals all the way to the
advanced through real world cyber security skills. So you will start with core concepts like operating systems, scripting, networking, cryptography and before moving into enterprise security, application security, malware analysis
and also ethical hacking. So what really makes this program stand out is the integration of AI throughout the curriculum. You will learn how AI is transforming uh both blue team and red team operations and also from automating
threat detection and response to understanding how uh attackers actually use AI which we all had a conversation about earlier and finally you will bring everything together through a portfolio project that simulates real world cyber
security scenarios. So by the end of the program you won't just have theoretical knowledge you will have practical experience and projects that can that interviews and that will probably give you that advantage uh while you appear
for all the roles um ahead and let's talk about the skills that are covered as a as a part of this program and I think one of the questions we often get is will I learn enough to be job ready and the answer is Yes. And this uh slide
will give you the reasons why. So this program actually covers over 21 in demand skills that employers are actively looking for. So you will build a very strong foundation in areas like network security, vulnerability
assessment, application uh security and also pen testing while also learning advanced topics like ethical hacking, thread detection, malware analysis and a lot more that you can see on the screen right now. What makes it specially
relevant today is the AI component. Also the discussion was around agentic AI. So you will learn how to use AI for thread detection and all the other things that we discussed earlier. By the end of this program you will have all the skills uh
applicable across cyber security roles. So you can be you can feel confident in So you can be you can feel confident in this field to deploy AI as well. And uh like Ha mentioned earlier. So now it is important that you have some AI skills
uh to clear cyber security roles as well. So if you're wondering what are some of the case studies um and the tools covered that you can now see that on the screen. So like I mentioned earlier as well uh you will have
hands-on learning um exposure and uh you will learn through real world case studies from leading organizations like OpenAI, Microsoft, Google, Amazon, lot more. You can see all of this in detail if you go to the course page. Uh
in the interest of time we will not be able to dive deep into any of these right now. But in the course page you will have detailed information about the modules about the topics that are covered and also your projects and uh we
also get a lot of question about the projects that are covered as a part of the program. Uh so these are some of the main projects that we are covering. So there is vulnerability assessment and pen testing on simulated web app. Then
there's digital foreign 6 recovering hidden data from JPG uh cyber intrusion meta spoiled and a lot more. So uh we will not uh if you want in-depth understanding of what these projects cover, you can definitely talk to a
learning advisor. So they can they will actually walk you through each module, what is covered. Uh and they will also tell you your best fit based on your current uh portfolio and where you're standing. Uh if you're a fresher, they
will be able to give the right advice for you uh and to also opt for this course and the benefits if you uh decide to go ahead with this. So um you might learn. There are a lot of other places which provide the same kind of courses
but there are a lot of reasons why we stand out. So you have the option uh for stand out. So you have the option uh for flex learn. So you will get uh uh you will have the flexibility to actually uh choose classes for weekdays or weekend
sessions. Uh you will have mentoring sessions. This I explained earlier as well. So you will have expert guidance from mentors for doubt clarifications and project assistance as well. Uh you will definitely have a cohort manager
for all of your queries. So we will ensure that there's somebody to help you throughout your learning journey and you always have that support uh support system with simply learn and we also have technical capability enhancement.
So you will uh learn from instructors who will equip you with hands-on skills are some of the highlights and we're really confident about what we provide.
You can definitely take a look at the course page. I will show you the QR code uh shortly. Uh yeah, so if you want to learn more about this in detail, you can scan this QR. It will take you to the course page. It has all the details in
depth. uh uh and uh for the course fee is 2 lak 32,000. You have the option to pay this in installments as low as 10,388 per month. So uh consider this an
investment. This could probably impact your career and land that help you land the job that you were aspiring to get. So sometimes it's important that we uh you know invest in learning uh especially with AI coming up in every
domain right now. So please go ahead and scan this QR code and like I mentioned if you want more details about this course uh we have learning advisor who can help you with the same. Uh so I'll quickly launch a poll if you would like
to know more about this course you can click on a yes and if you're interested to enroll you can click on a yes uh and our advisers will reach out to you at the earliest. So please give me a minute and I will um quickly share the poll
with all of you. Uh Shua, sorry to interrupt. Meanwhile, shall I take one the Q&A which I feel >> worth asking? >> Yeah, absolutely. >> So, one of the learners is asking um so
what is your opinion about um a GRC as a career? Is it good for freshers or not? >> So, basically that's a very good questions to good question to ask. So as
far as you know GRC is basically making sure the company follows the rules and manages the risk properly right so thinking like complaints u complaints thinking like complaints u complaints things and audits uh and policies etc.
So it's a great option if you like tech but don't want to code all the day. Okay. So you should start with basics like some certifications will be there which were um like uh which were something like ISO27 something is there
as far as I know. Um and also Cisco related certificates also there. CISA certification is there and try to understand the cloud basics also since the companies want GRC people who actually get the tech. Okay. Of course
the pay like the compensation is solid and the work life balance is also better and the work life balance is also better than pure engineering roles. So yeah that is what I can tell and also some people are asking um most of the L1
people are asking um most of the L1 operations are being um uh replay like being automated because it's kind of repetitive tasks. So of course most of the L1 tasks are repetitive like checking alerts, gathering info,
following checklist, right? So AI is really good at that part. Now it can it can look at the alert. It can uh pull the related information together and the related information together and tell if it's serious or just something
which can be ignored. So a lot of the basic task is already being automated. But AI still struggles with judgment. Okay. like understanding um is this a big deal or not? Is this going to affect the companies company systems or not or
explaining things to people and handling tricky situations. So humans are still needed for that. So what's really happening is companies need fewer people happening is companies need fewer people just watching alerts all day but they
need they still need a people to review that what AI flags handle tricky cases that what AI flags handle tricky cases and improve that AI system itself. So, so my honest advice is if someone is starting out in SOC L1 roles, don't just
plan to be an alert checker long term. Uh try to go fast into things like um I can say detection engineering and we have uh risk analysis, root cause have uh risk analysis, root cause analysis or handling complex incidents.
Okay. So yeah, that's what I can actually um tell you. So yeah and also actually um tell you. So yeah and also somebody were asking me um as an S sur
have you seen agent agentic AI help reduce alert or improve MTR in uh real reduce alert or improve MTR in uh real real what I can say real environment like real production environments. Yes mainly in a triage and diagnosis stages
mainly in a triage and diagnosis stages but not full remediation. Okay. So of course u when something breaks a big chunk of time goes into understanding what went wrong. But here AI can quickly check the logs recent changes change
management databases and the past similar issues and give a quick summary what happened and what I can do which saves a lot of time and manual digging. It also cuts down alert noise. Instead of getting 20 alerts for the same issue,
what it will do? It will group all the same alerts into one single alert. So it will it will only trigger only one single alert that harsh this specific service is having some security issues. Check. Okay. So for actually fixing
Check. Okay. So for actually fixing things, AI is allowed only to do safe simple actions like uh restarting a very small service or maybe uh clearing the caches something like that. So anything still needs a person to approve it. So
overall this can cut the resolution time by 30 to 35 or maybe 50% also sometimes for common issues. Not because of AI replace engineers but um because it
removes the slow repetitive and manual manual work. Okay. So that's what I can tell. >> Thank you for answering all of their questions Hara. I can see a lot more questions about the course. Again,
will not be able to dive deep into the course details. So, I hope all of you have responded with a yes here. If not, you if you still have questions, I I saw some questions about CISSP. Uh you can go to the uh course page and there will
be an option to talk to the advisor. You click on that and our team will be there to help you. So, I'm going to quickly end this poll uh right now. I really hope to see a lot of you as a part of Simply Learn's journey. Uh and with that
Simply Learn's journey. Uh and with that I will also p uh you know uh I'll launch the next poll which is the certificate poll. Uh so all of you please share your full names as you want in the certificate. Uh we already have captured
your email ids. Uh so you will be receiving your certificate the slide deck and the recording to your email ids within the next 48 working hours. So if you don't receive it within that time then you can write to us at
webinars@simplearn.net. net and if the this poll is not visible to a few of you again you can write to us at webinars@simplearn.net net. I'll just quickly write that uh in the chat for your reference if you guys are
facing any issues in accessing this whole data. your responses I think I will quickly take this time uh to thank Harsha for
such a wonderful uh session. It was great hearing from you learned a lot. It was really insightful to be a part of this webinar for me as a host and I'm asked. So I'm hoping that it was really impactful for all of them as well. And
uh before we wrap up, I always ask uh the guest speakers if they would like to leave the audience with any one piece of advice uh before we wrap up. So uh would you like to share your thoughts on that? >> Sure. Sure. So I really had a great time
this evening postwork today. I'm having a very uh stressful day at maybe I'm uh handling some monitoring things but now my habit is talking to students and maybe listening to them talking to them makes my day very good. So thanks to
everyone for your time and thanks Suita for inviting me for this session and and making me a part of this journey. And when it comes to advice, so I simply follow
um I think it's relevant. It's it's redundant or it's everybody will say redundant or it's everybody will say this but consistency is the key. I also felt the same when my seniors told that. But later I realized consistency is key
But later I realized consistency is key not only for this. I built a gym habit. I built a lot of habits with consistency. So build a small atomic habits. Let's say you are going to learn agentic AI. So set an habit of reading
or doing hands-on about agentic AI every day for 10 minutes. So let's say um habits works like this. You need not to get 100% productivity over it. It's just
showing up. Let's say you just open laptop and read for 5 minutes. That's completely fine. And second day you open laptop and just read for 10 minutes. That's completely fine. So here it's just showing up. It builds the habit.
And also very important thing is if you miss today, don't miss tomorrow. So here one day is emergency but two days creates a pattern of missing the habits. creates a pattern of missing the habits. So any whatever it happens I do follow
the same for my gym activity. I do follow the same for my all day activities. may be drinking 4 L of water. One day I I might miss 4 L of water but I can at least drink 3 L of water. So here it's just that to not
break your habits just make it at least something to happen. Okay, it's not only about this cyber security or anything anything in our daily life just creating small atomic habits checking them every day with at least small small wins
celebrating small small wins makes a lot of difference. Okay. So yeah that's it of difference. Okay. So yeah that's it that is a simple advice I can give and keep track of habits don't miss two days
miss one day but don't miss second day that creates a pattern that is very very I am telling you as an experienced person because just missing second day as my gym I didn't went to gym for two months so don't do that okay don't do
>> thank you so much Harsha for that wonderful advice I think it applies to including me struggle with this consistency. So hoping to build that consistent. Uh with that I'm going to end this poll right now. It was
wonderful to have you guys also. I think uh it was it's always a good webinar depends on the kind of audience we have and today you guys were amazing. Thank you all for staying back. I know a lot of you would have had a long day uh at
work and you still came and attended this webinar. So thank you so much. I simply learn or at least for our upcoming webinars. I wish you all all the very best in whatever career that you choose and hope you all have a
wonderful day tomorrow. So, thank you so much everybody. I'm going to uh end this much everybody. I'm going to uh end this webinar right