[0:00] what if you could turn your website in
[0:01] all of those PDF documents manuals
[0:04] guides uh ebooks and reports into a
[0:07] helpful chatbot imagine asking questions
[0:10] and getting instant answers directly
[0:12] from your website your documents or any
[0:15] ebook will guess what I found a way to
[0:17] do this in less than 10 minutes in this
[0:20] video I'll show you how easy it is to
[0:23] build your own AI chatbot with custom
[0:25] knowledge base without any coding skills
[0:28] you can do this for your business or
[0:30] other businesses and start making money
[0:33] by the end of this tutorial video not
[0:34] only you'll be able to create a chatbot
[0:37] and train your AI based on your own data
[0:40] using PDF HTML txt CSV Word documents
[0:45] and PowerPoint presentation files but
[0:47] also integrate the chatbot into your
[0:49] website Facebook Messenger Flack Discord
[0:53] or Telegram in this video you learn how
[0:55] to create a data story agent how to add
[0:57] add structure data to your data agent
[1:00] data store how to enable voice and chat
[1:03] experiences for your agent how to test
[1:05] agent and simulate customer questions
[1:07] how to V conversation history and
[1:09] analytics for your agent all of that in
[1:12] about 10 minutes without any further Ado
[1:15] let's get
[1:20] [Music]
[1:24] started all right so before we start let
[1:27] me show you the document that we'll be
[1:29] using thing as the data source this is
[1:32] budget of the US government and it's 188
[1:35] pages long so to get started we'll go to
[1:37] cloud.google.com
[1:38] uh log in with your account and click on
[1:42] console from the console click search
[1:45] for agent
[1:48] Builder click on agent Builder before
[1:52] creating a new app make sure that the
[1:54] vertex AI API and the dialog flow apis
[1:57] are enabled if they're not enabled get a
[2:00] prompt and then you can enable it on the
[2:04] screen select
[2:06] chat and provide the company
[2:10] name provide the agent
[2:18] name click
[2:21] continue we'll create a data store this
[2:24] is where we can create a data store
[2:26] click on create data store you get three
[2:29] options
[2:30] here you can provide a website URL and
[2:33] it will automatically crawl website
[2:35] content from a list of domains that you
[2:36] defined you can provide the data from
[2:40] your storage bucket or you can import
[2:42] data manually by calling the
[2:44] API I'll select the cloud storage and um
[2:49] you can import a folder or you can
[2:50] import a file I'll select file and head
[2:54] browse I created this bucket earlier
[2:57] where I uploaded that PDF document that
[3:00] I showed you uh I've already uploaded
[3:02] uploaded it um but for your case you can
[3:05] upload any document to your uh cloud
[3:08] storage and you'll be able to select it
[3:10] from here click
[3:12] select so PDF html text files CSV Word
[3:17] document and presentation files fall
[3:19] under the unstructured documents now you
[3:21] can use that as the data source or you
[3:24] can use Json L for unstructured
[3:26] documents with metadata information and
[3:28] click on this link for more information
[3:30] or you can use CSV file for structured
[3:33] um frequently Asked data this is only
[3:35] for chat but for our example we're going
[3:38] to be using the unstructured document
[3:40] click on continue I'm going to name the
[3:43] data store us um budget 2025 25 and
[3:50] click on create it's going to create the
[3:52] data store okay so it created a data
[3:54] store I will select the data store and
[3:57] click on create
[4:11] okay so the data store has been created
[4:12] over here I can click on over um on this
[4:16] link uh to see more
[4:19] information um on this screen I can see
[4:22] the
[4:25] activity and I need to make sure that
[4:27] data has been imported successfully with
[4:29] without any
[4:32] errors and we are using the digital
[4:34] parser now we can hit um preview to see
[4:39] the
[4:45] preview uh now we will test our agent uh
[4:48] testing is an important part of working
[4:49] with conversational agents to find box
[4:51] identify limitations and simulate a
[4:53] customer experience you can start an
[4:55] interactive session with your chat bot
[4:56] to see how it responds to various
[4:57] questions that a customer might ask
[4:59] click on the test agent Link at the top
[5:02] right hand
[5:03] corner um this is where we can uh test
[5:06] the chat bot we can write
[5:10] hello and we have more control over
[5:13] customizing the response of the agent
[5:15] from this screen um we can enable
[5:19] logging um by going here and checking
[5:23] this enable conversational history by
[5:25] checking this box uh we will get the
[5:27] conversation history uh let's suppose we
[5:30] put this on our website and people
[5:33] interact with our chatbot and if you
[5:35] want to be able to see those
[5:36] interactions you need to check this box
[5:39] and hit save we click on manage click on
[5:42] integration let me show you a very cool
[5:44] thing uh you can add a phone gateway to
[5:46] your Bot which will make use of speech
[5:48] to text and text to speech capabilities
[5:50] in the Google Cloud you can head on this
[5:53] manage uh button to enable that feature
[5:57] but for now I'm happy with the text
[6:00] uh chatbot and text based I can click on
[6:04] connect we can enable this
[6:06] unauthenticated API which means that
[6:07] will allow people without authentication
[6:10] to interact with the chat bot and I'll
[6:13] click on enable an authenticated API
[6:16] this is the code that you can use to put
[6:18] on your website and you'll get the
[6:21] chatbot on your website so to see how it
[6:25] would look like click on try it now and
[6:27] you get this um you get this icon at the
[6:30] bottom uh right hand corner so you click
[6:33] on this and this is the chat box that
[6:36] you you can put on your website what are
[6:41] president's top
[6:43] priorities for year
[6:49] 2025 the chatbot read the document and
[6:53] responded with present stop economic
[6:55] prorities lowering cost for American
[6:56] families the administration is working
[6:58] to bring down the uh prescription drug
[7:00] cost health insurance premiums utility
[7:02] bills cost for everyday Goods all while
[7:04] taking on junk fee studs some Airlines
[7:07] Banks and other companies use to rip off
[7:09] Americans and also it's grounded the
[7:12] data and if I click on this link it will
[7:14] show me exactly where it found that
[7:16] information from and as you can see it's
[7:19] on page nine and this is where the
[7:21] chatbot found that information is
[7:24] lowering cost for hardworking Americans
[7:26] the administration is working to bring
[7:27] down prescription dve cost health
[7:29] Assurance we're we are grounding it with
[7:31] our own data source so let me ask it
[7:34] another question let me ask what has
[7:38] president invested
[7:50] in the president has invested in America
[7:53] and American people by signing into law
[7:56] historic legislation such as the
[7:58] American Rescue plan the bipartisan
[8:01] infrastructure law the chips uh and
[8:03] science Act and the inflation reduction
[8:05] act let me ask it another question like
[8:08] um what are some of the
[8:12] accomplishment of the Biden
[8:19] Administration the bid Administration
[8:21] has made historic Investments and mental
[8:23] healthc care including nearly 1 billion
[8:25] to support the 988 suicide and crisis
[8:28] Lifeline the administration is also
[8:29] committed to investing in training more
[8:31] mental health professionals in his
[8:33] expanded certified community behavioral
[8:35] health clinics Nationwide when testing
[8:37] your virtual agent if you find that your
[8:40] bot's answers do not meet your
[8:41] expectations you can always add more
[8:44] URLs to the data source to improve how
[8:46] your Bot handles questions and answers
[8:48] you can add more URLs it can add more
[8:51] documents um other than testing many
[8:53] different paths the customer might make
[8:55] how else can you determine where the
[8:57] problem in friction points or in your
[8:59] version ual agent conversation history
[9:01] in analytics help uh so we enabled I
[9:05] showed you where you can enable that uh
[9:07] conversation history under the agents I
[9:10] already enabled that and to get to that
[9:12] conversation history you can click on
[9:14] conversation history and this is where
[9:16] you're going to see the conversation
[9:17] history and I can click on any of these
[9:20] uh and it will show me uh my interaction
[9:24] with uh the chat bot I've asked it's
[9:28] going to show the history of my
[9:29] conversation and once you put that on
[9:31] your website and you would be able to
[9:33] see the interaction of your chat bot
[9:35] with your visitors website visitors
[9:37] congratulations using the verx AI
[9:39] conversation in dialog flow CX you have
[9:42] successfully created a data store agent
[9:44] your virtual agent can answer hundreds
[9:46] of different questions about your
[9:48] website or any document PDF manual
[9:52] guides ebooks and reports at sector that
[9:54] you may have you don't have to go
[9:56] through the manual process of creating a
[9:58] large number of intense training phrases
[10:01] response messages we literally created
[10:03] this in less than 10 minutes this video
[10:06] was requested in the comment section by
[10:08] a few of my subscribers I read all of
[10:10] the comments and i' love to hear your
[10:13] thoughts um feel free to share your
[10:15] ideas for upcoming videos if you don't
[10:17] have any questions or feedback you can
[10:20] support my work by just typing coding
[10:22] money in the comment section it really
[10:25] helps with the YouTube algorithms if you
[10:27] found this video useful a thumbs up
[10:29] would be highly appreciated stay tuned
[10:31] for exciting new videos to ensure you
[10:34] don't miss out please subscribe and hit
[10:36] the notification Bell icon you may also
[10:38] want to check out the video that pops up
[10:40] on the screen somewhere until next time
[10:44] thank you for watching