Build a FREE AI Chatbot: 24/7 chatbot that answers questions, resolves issues, and nurtures leads
Let's build with Zapier and AI (100+ videos)📅 2024-03-25
Transcript
[00:00] you may know already but zapier has the ability to create chat Bots specifically these chat Bots can be integrated into your website or alternatively be used internally for your business zapier announced a new feature that now we can actually use tables as a data source for our chat bots so in today's video we're going to learn how relevant this is to us for our business on top of that we're going to understand what this means in the context of internal use for our business let's jump in so I've done a good amount of videos on zapier's feature specifically when it comes to chat Bots and this is different from zap
[00:30] Central zier Central seems more like very much internal use plus like next level when it comes to automations and communication in that context zapier chat Bots if you're not familiar with it think of it just like a regular old chatbot but zapier's version as it allows us to connect to their entire ecosystem when it comes to interfaces and automations I'm going to leave this video right here which is just a video showing you how to integrate it into your website if you're interested in doing that but in today's video I'm just going to go over the new feature that got added recently which is pretty cool which allows us to Cipher through table
[01:00] data this can be used in the context of either your chatbot on your website EG you have a product CSV you want to import into tables and then chat with the chatbot or alternatively use for internal data like looking through data that you have internally and talking to a UI that is like a chat likee interface let's see what we can do now so I got some data here I'm actually going to add a couple more rows so bear with me so I'm going add you know a couple more records here so we can actually prove and gut check this new technology and then let me add some names here say Sam
[01:30] Adam Ashley I'm assuming that you would already have data here Jennifer and then we can go ahead and add some emails here basically the major use case of this feature is that if you are leaning into zapier heavily right now and you're kind of already building on top of its ecosystem when it comes to interfaces think of that like landing pages and maybe in the future like websites and then tables think of it like a database that you can start leveraging so maybe a more active Excel sheet and then obviously the Big Kahuna here which is going to be the automations itself if you are leveraging all of it this is
[02:02] just like another tack on where you're like okay cool now I can actually look through my data faster and in a more expediated way and for the numbers we'll just go ahead and say everyone has the same number we're going to go ahead and change some variables over here we'll go ahead and do 5100 100 500 500 plus and 1050 because the purpose of this video is I want to gut check this I want to see how good and how accurate this can be when searching through our data got my table set up let's goad and integrate it into
[02:32] our chatbot coming over to my chatbot here I have named it chatbot data reader seems like an obvious name because it's going to read data for the greeting typ I'm going to say static greeny text fine ask input placeholder ask me anything that's the thing right here we've done a bunch of videos on this y'all so you can check them out open AI authentication I can leave as none for now and then the model you want to choose really is dependent on your use case so because of the fact that we're searching through D tables we're searching through a bunch of data and the the EG the use case here is just
[03:02] finding and basically formatting data we can stick with 3.5 because of the fact that this model actually is pretty effective in the context of formatting and finding data if we were using this chatbot in a more higher level threshold when it comes to the types of outputs we're looking for we would increase it to four what do you mean by that Corbin if you were going to use this chap in the context of running social media captions writing articles whatever it may be the end point or the end value point is a human will look at it will visually look at it then you want to
[03:33] increase to four but if we're just searching through data and ciphering through data we can proceed with 3.5 creativity we can leave it at 7 if you want consistent outputs we'll leave it at 0.1 if we want it to go a little crazy on us let's go to one or random 0.1 let's leave it at 6 actually we'll go in the middle 0. five disclosure none of this is really relevant in today's use case because I just want to show you if you were to use this internally on your team we H instructions first we're going to hit save all right now that we'll be saved
[04:04] here instructions and this is not too relevant here we can just you are chatot you're helpful assistant we could basically laser this in a little bit more for the context of what we want to use it for so maybe we'll say you are a chatbot helpful assistant when looking through our customer data it's going get more advanced save changes here we're going to add a source The Source we're going to add today is going to be tables this is going to be a big deal y'all here's why it's going to be a big deal up to this point we've only been able to add a file and that file has only been
[04:34] one megabyte and as you'll if you watch that video I referenced in the beginning you'll notice that we will have to basically provide that file in in a text format so that the actual file size is as less as possible does that make sense yeah that makes sense therefore so it allows us to upload it what's great about this feature and allowing us to go to tables is we can deal with a lot more data that we can reference we do a thousand rows that's powerful we're going to go and find the table it's going to be customer list we can add a description if we choose to do so just say add Source it's going to load it in
[05:05] or it's going to fail to connect the data source it's going to load it in sometimes software just requires a real quick refresh don't worry about it session storage local storage is it cashed is it not cashed do you want to eat cookies no okay here we go so we got when results are not found generate an AI response about the knowledge Source or we can set a custom message here so for example sorry I'm not even able to provide the answer that could be useful so we don't just get randomness we're going to go ahead and say anytime you make a new change to your data source specifically in this context a table you're going to hit sync but since we
[05:36] don't need to do that for now we can start testing this out so let's go and gut check this first off how many customers do we have and it should say five I'm going to be honest with y'all I was actually running into a lot of Errors too many errors so I was like what's going on here obviously I didn't do something correct and this is what I didn't do correct in your instructions you have to put in very specific dictation to reference the specific knowledge Source you wanted to use in the chat if you don't do this it's not going to give you any answers it's actually going to be like very confused
[06:07] therefore use a line like this I reference the table by saying custom list and in the actual pop out I say use customer list knowledge source to answer questions let's open this link right here and proceed furthermore as you saw off that link you can actually share this with your team so the use case here would be like you upload your table that you're already using with zapier and now you can kind of Proctor it in a manner like this what are all our customer names we should get five here there we go I'm actually going to shrink y'all super small from here let's G check a
[06:39] little bit we're going to say give me the emails for Adam and Corbin see Adam gmail.com we should see cor or web Rista at webcafe webcafe cameras.com there we go so far so good give me all the customers who were interested in our service let's see if it's smart like this so what I'm gut checking here is the check boxes here so we should see Corbin Adam and Jennifer Corbin Jennifer and Adam okay next thing to check out
[07:12] here is let's see if it can identify based off company size and understand a variable like this so we're going to say give me the individuals who have employees or the amount of employees is over 100 which customers have a business with employees over a 100 should be two can I do this okay Adam Ashley let's see Adam yes Ashley yes pass that now at a service level you might be like Corbin this is what's the point here you have five different customers why would you
[07:44] need to talk to a chatbot like interface in this context well if you remember when I imported this table it can do up to a th000 rows therefore this kind of conversational setting here would be a lot more effective in the context of like needing to Cipher three a thousand different customers for specific context and for specific data points so much so I can go as far and say for Corbin what were our additional notes so maybe I'm ciphering through customers I'm like you know what wait
[08:15] what was Corbin again okay these were the additional notes we had okay that's a new integration between zap your tables and chat Bots let me know if you're going to leverage this in your business I see this more along the lines that if you're already using zapp's ecosystem with interfaces chat Bots and tables this just seems like a like a cherry on top where it's like okay well now I can actually just lean in further into the ecosystem and leverage this kind of Technology also zap your central is crazy if you don't even know what I'm talking about you should check that out it's a whole another part of the ecosystem that's taking this stuff to like the next level when it comes to Ai
[08:46] and automation but without further Ado I'm going to leave a playlist at the end here you should check out I'll see you in the next video that's a playlist I was referring to this is going over Ai and zier and everything you need to know that's a random video of my face maybe not my face and this is my face click it don't click it may or may not happen