Zapier and ChatGPT For Google Docs Analysis | Tutorial
Let's build with Zapier and AI (100+ videos)📅 2023-06-03
Transcript
[00:00] welcome back to webcafe AI we do daily kgbt and AI videos for your personal and business life in today's video I want to show you an AI automation that allows you to take a lot of text and then analyze it quickly with AI all right let's go ahead and begin in order to do that we're going to hit create here and do news app essentially the idea behind this app is that you can take a lot of text maybe from illegal contract a resume uh some type of bulky PDF you can paste it into a Google doc and then we'll use gbt to automatically analyze it for different parameters that
[00:30] you're looking for specifically I'm just gonna go ahead and rename this to text analyzer all right coming over here we're going to add the trigger of Google docs and essentially the trigger in this use case we're going to go ahead and say when a new document is created in a folder so we have more specification here and we're going to hit continue here we're gonna go ahead and choose our courses account as always these apps built in these
[01:02] tutorials will be found in the description below and then we're going to go ahead and find the folder that put in I put it into the YouTube zaps and then just for your reference I'm gonna jump over real quick and as you see here this is going to be the example uh piece of text that I'm going to be analyzing I just grabbed the random resume online and I paste it in here go ahead and use the this zap in whatever context that you plan on using it for maybe you have like a 10 page legal contract that you want to paste in here and then have the automation occur but from here we're going to go ahead
[01:33] and hit continue we're going to test that trigger all right let's go ahead and confirm that it's on the correct one so in order to do so I'm going to type in the file name which is going to be resume template all right perfect it has found the correct one from here we're going to go ahead and do a couple things so the first thing we're going to do is we're gonna do a gbt block and we're gonna do the event of conversation from here we're going to go ahead and you have like multiple options here for smaller files I'd suggest you to do the execution of what your intention is
[02:04] right off the bat but if you don't have smaller files here go ahead and start with a gbt 3.5 Block in order to format the text better so we're going to go ahead and say based on this and then get context of whatever you're trying to summarize here so in this context we're summarizing a resume based on this resume if you send me one parentheses and then we're going to do the file raw plain text here and then we're gonna do parentheses again we're going to say uh please generate a summarization
[02:35] of these six main points and then we're gonna go ahead use 3.5 as this is more grunt work we're going to use a memory key of res out and then we're going to continue here and we're going to test this action okay so based off this lengthy resume it was able to print out these six main points basically the context of why I made this block was more of think of it as like Assurance so if you guys down there essentially put in a lengthy legal
[03:05] contract or a piece of text file that's very large you're going to want to set up a block like this that essentially summarizes the data to the main point so maybe my intention here was wrong maybe what I'm really looking for here is not necessarily the six main points maybe I'm looking for please generate a summarization of work experience specifically what's occurred within 2020 and onward so maybe I want to get very specific on this type of summarization if we're dealing with that
[03:36] lengthy of a file so I'm going to reset this again to show y'all and then once we get that output I'm gonna go ahead and proceed to the next stage of this as if I didn't need to summarize the dog like this and I had you know not too much data to mess around with all right so as you see below me we are getting access to specifically what I asked for when it came to work experience past 2020 so since 2020 Jason Miller has worked at Amazon warehouse Associated Miami Gardens when he well performed all Warehouse labor duties such as parking picking counting and so on all right so
[04:06] everyone had to rename that block to compressor the purpose of this block as I said before is really to you know make your data smaller if you're familiar with zapier and AI you may get that really pesky air there's just 504 error that typically means and is correlated with too much data being inputted into the conversation therefore it times out so use this as a trick for any other context that you're dealing with that kind of error sometimes you just gotta mitigate and minimize the amount of data that's going into the initial input but we're going to proceed now as if we didn't have to memorize and minimize the
[04:37] data so I'm going to ignore this insurance block here I'm gonna create another chat DBT lock and I'm going to go ahead and say conversation so assuming that we don't have too much data when we're dealing for initial input here we can just do the same thing and give context and say based on this resume semicolon parentheses document file plain text here we can do shift enter enter I went ahead and generated some main points that maybe a job recruiter would look for I'm going to say generate
[05:09] what's applicable here let me do semicolon and then essentially I generated four things so we got relevant scales job specific skills and competencies work experience education accomplishments and then let's just say the role is um Marketing Manager once you have provided the previous summary give me a score of 0 to 100 100 being
[05:44] very qualified of the candidate so as you see here this can get pretty complex but for the purposes of this tutorial I really just want to show you how to parse data and have data or basically have the ability for AI to analyze your data automatically so from here I'm going to go ahead and make sure we up our model to gbt4 as we're gonna deal with a lot more comprehension in this context and everything else can contain the same if we want to maintain fixed structure on the output I'm gonna go ahead and just say out for from here
[06:15] I'm just going to continue and then test that action okay so as you see we got our output here relevant skills experience picking packing counting and record keeping work experience and as we keep going down here we get to our final score here at 40 out of 100 although Jason Miller has demonstrated strong skills and organization time management and Warehouse operation he lacks direct experience for marketing management as a result the candidate receives a score of 40 out of 100 indicating he may not be the most suitable choice for marketing manager
[06:46] role as you see here this can come into context with a lot of other things when it comes to analyzing text in different Realms and different areas of business but in this specific context and this could be a very useful tool to maybe uh pre-text or check out resumes before they you know really take into consideration or go to a first form interview now if you feel like you've learned something at this point make sure to like the video it helps us here at web Cafe Ai and it's completely free but one other thing I want to show y'all is this insurance block and how it connects to this specific block which would be more of the output so I'm gonna
[07:17] go ahead and rename this to Output in this context essentially let's say you input the original data into your output block here and then you're getting that 504 error well what you can do now is we can use the compressor and we can go ahead and grab the assistant response message and now that we have the compressed version of that data we could probably come to the same consensus consensus post-generation okay so I know I said that was the last thing on this specific tutorial but actually I do want to show you one last thing just in case you want to make you know the
[07:47] final little bow to this whole flow for example let's say you made a new document you put in our resume in this context you can always send it as an email like this making a Gmail block you can send an email and then obviously just choose your account so for us to be courses hit uh continue here and then all you need to do is send the email or put in the fixed variable address or whatever address you want to send the email to in the to section and then you would just provide a subject line like resume overview and then in the body we would
[08:18] just put in the output that we just created there and from there you're just a continue publish and you'd be good to go so that does it for this tutorial as always you can find the zap down and below if you're interested in learning more about how Chad gbt and zapier can be leveraged make sure to check out the playlist at the end of this video as it goes over a ton of AI automation as that is the main focus of this channel but without further Ado I'll see you in the next video thanks for tuning in and yes surprise I'm an AI Avatar make sure to explore more here at web Cafe where we demystify
[08:51] AI for your personal and business life until next time