Lesson 4 - Zapier: Folders, Flows, and Functionalities

Start an AI Automation Business📅 2024-07-31

Watch on YouTube ↗

Transcript

[00:00] welcome to module 2 building and a automation agency where we're going to go in depth with every single piece of software that we outlined earlier and we're going to go ahead and start off this module with zapier and show everything that's fundamental about this platform when it comes to building out an AI automation agency to start off let's talk about pricing so this platform ranges from free all the way up to a company level backend if you are just starting out with a Automation and you want to get more comfortable with the platform start start with a free

[00:31] account as you're going to be able to run some automations and really get an idea of the potential that you can do with zapier now there's a couple things I would like to point out first being task Associated of each plan in this context just so you know task are the blocks or steps found within an automation for example a chat GPT block publish an Instagram post create a tweet that counts as one task what's important for you to know is that tasks are going to be expended during a live autom and during a draft automation that means

[01:03] when you're building out a Automation in the back end and you expend a test that counts as one task therefore as you keep practicing and building out within the backend you will be expending tasks so there is going to be a high probability you'll probably need to upgrade to the next plan when you upgrade your plan in zapier it has two main benefits first one being it will have a different perk that's unlocked and the second being the underlying cost of the zap your task is going to be less now in the context that you build out an a automation with a recurring method

[01:35] where you have the a automations in your back end and running it for a company you're going to be able to get your zap your task to very very cheap levels for your reference let's go ahead and draw out a picture here we're going to start with let's just say you have a starter plan and the cost associated with a task in a starter plan we're going to go ahead and put in 30 divided by 750 task and as you see there it's going to be around 4 cents per task now assuming you're doing a lot of testing or you want to move up your plan for XYZ reason

[02:07] let's go ahead and see how much it would cost for let's say 2,000 task at $73.50 all right so we're going to go ahead and put 7350 here divide that by 2,000 and there you go we have each task being 0 3675 rather than the 4 cents so as you scale and you get larger plans the underl cost your a automations will be less and on top of that the task per usage will be less now that you understand that the next thing you need

[02:38] to understand about each plan is the perks you unlock as you see here the context of an AI automation agency probably the more fundamental perks that you really need to have unlocked are going to be the multi-steps filters and formatters and versions the perks unlocked for professional are still important such as unlocking premium apps but are not as fundamental as the perks that are locked in the this first tier called starter now one big thing I'd like to point out is the first three tiers are going to be associated with an individual while the last two tiers here

[03:10] are going to be more company oriented so if you have multiple individuals and as you see here it's kind of outlined here with users so we got one one one and then unlimited uh I will say that the cost Associated as you see here is a little bit more relative to the plan even though you get the same thing but just more users in the team version okay so now that we understand the plans and the costs associated with running a zappier account let's go ahead and go over everything you need to know about the platform from start to finish now when just looking at the UI here you'll

[03:42] see your profile up here you'll see the ability to contact sales up here an ability to search your relative ABS here usually we don't have this here this is just a notification that they're doing a little bit of a user conference then we have our Beta app here right here you can see we have recent zaps that you working on then you got recommended zaps from the community and then just you know different information down here uh relevant to something maybe you're interested in such as joining the Early Access program okay so now that we got an idea of our dashboard let's go to get

[04:14] understanding of all the different tabs that are found here starting with zaps now one really cool thing about the zapier platform is our ability to organize our zaps due to the fact that as you grow you're going to get a lot of different zaps for a lot of different reasons as you see here that is why having the ability to create folders is so fundamental so in the context of running an a automation agency here are the top folders I suggest you to create right off the bat the first folder here we can just call backend now the purpose

[04:46] of this folder should be used for all AI automations or just automations in general that are utilized for your own business think of this as your business's backend in the sense of anything that you need to get aom ated and done at an expediated rate for your personal business you should be building it out within this folder here just to give you an example here if an individual went ahead and subscribed to your newsletter or some type of marketing material there is a really cool zapier flow you could run where essentially anytime an individual

[05:17] provides their email for a marketing service you can automatically add that to a Google ads campaign group there are plenty more automations that will come up over time when you're running your a automation agency but use the back back end folder as a way just to organize every single zap year that is associated with your business operations the next folder you should be creating here is going to be called clients now depending on how many clients you currently have you should be basically organizing it off that so if you have a lot of clients You' probably

[05:48] do some type of alphabetical system here but if you're just starting out you can typically just do a client folder and what's great about zapier is we're able to create a folder within a folder H these ellipses here we could hit new folder here and let's say the company is called Apple's Co we can simply do Apple's Co here and then now we have a folder within a folder we can click that click on that and as you see here we're in clients and then we're even further in clients and it's Apple's Co so any a automations that we may be working for

[06:19] on this company or already have built out we can have placed in this folder for organizational purposes now when you're running an a automation agency what you'll notice is a lot of times you may be experimenting or developing a automation that isn't necessarily associated with a client but it's more for testing to see the capabilities of zapier and AI on this platform and essentially the purpose of beta is to create AI automations or automations in general that are pushing your limits to really understand the capabilities of this platform and maybe for the ability

[06:52] to provide a specific service that a business had requested but you're not sure if you're able to do yet perfect now you really know how to organize your zaps once you start building them out for potential clients or for the back end of your business let's go ahead and just create a zap and walk through the platform of the zap now if you're part of the Early Access program you'll be able to change between classic editor and visual editor there is no fundamental difference between the two it's just a different UI for you to play around with this being the classic editor and

[07:23] then if we come over here this being the visual editor personally I find using the visual editor for more complex flows to be more effective but for most things you can still use the classic editor following this as you see at the start of every single zap we have a little thing suggesting you know inputting some text here and getting a possible AI outputs on what the flow could look like personally I suggest not to use this feature as it can be misleading and take you down the wrong rabbit hole because

[07:54] of the fact that it isn't really been fine-tuned that right yet therefore in the future once this feature gets better I could see it being very powerful in zapier but for right now I would ignore this beta feature right here now if you come up here you'll see that you just change the zap here so we just say zap name that's where you're going to be able to change and organize your names here if you come over here you get a lot of the different options that are associated with each zap starting with status status tells us whether the zap is active or not and if there's errors

[08:25] within the zap and outline is really cool here this is going to give us basically how the entire zap looks in a different format if you come over to settings we can add a description we can add it to a specific folder we can choose our time zone versions if you have this high of a plan allows you to roll back to specific versions of your zap if you maybe go down a rabbit hole that isn't effective for what you're trying to do history gives you a history of the zap in every single one of these runs

[08:55] this is going to be important when coming into cost and essentially you know usage in the future details you get a different type of UI look of the underlying zap here with the overview tested usage zap runs and change history clicking help gives you the ability to ask to zap your community if you hit a wall or you can check out the help center and then finally one of the newest features added to zap your flows is the ability to add notes um essentially here's the context uh if I add a scheduler here it now allows me to

[09:26] add a note here so maybe there's something I want to remember the next time I come back to this automation um make sure to schedule on Thursday so maybe I need to leave that note because a client told me that and I don't want to forget okay I saved it I got to go do something else right now this is a good way just to save specific notes for specific blocks now that we understand the side tab here we can go ahead and minimize it if we choose to do so we now need to understand the actual just automation flow so if we come over here to change

[09:56] we have essentially the most important part of any flow which is going to be your trigger this is what's going to cause the entire automation to run so it's very important that you know what you're doing when you set up your trigger because this essentially is you know like the first Domino getting hit and all the other dominoes follow with triggers it really depends on the context of what you're trying to achieve some apps have the ability to do triggers and actions some apps you can only do actions it really depends on the app type and the what you plan on

[10:27] essentially doing for example I can't trigger off a path because a path is later in a flow and it's basing its decisions off previous data for the purpose of showing you an automation we're just going to go with the trigger of schedule also note that these are native apps built by zapier and there's 20 of them or over 20 of them these are powerful these are going to give you a lot of abilities when manipulating the underlying data that we get with AI so it's good to know I mean some of the ones that stand out to me right off the bat it's going be filter formatter we

[10:59] hook RSS paths um looping is a really cool feature there's a ton of stuff we can do with the natively integrated apps from zapier moving on from this though we're going to go ahead and just choose schedule now when you choose the trigger you'll be prompted with events and essentially events are the different API calls that the app offers what I mean by that is essentially let's say you know let me go ahead and choose a different app here we're going to do Twitter

[11:30] essentially when I hit the event here and you see liked my tweet my tweet new follower new follower of me search and Geo all of these are formatted in a really nice UI that are very clear to us what it means but in the context of what's happening in the back end essentially zapier is grabbing different API calls in order to do these type of tasks and grab this kind of data so what that means is that some apps maybe have more trigger functions but aren't readily available Within zappier to clarify on that point

[12:03] essentially what I mean is that there is some context where maybe an app has a very specific trigger you don't find it here you may need to do a web hook where essentially you set up a custom API call if this sounds confusing don't worry we do an entire video found in this course of how to handle web Hooks and ways to call for custom API calls I mean this is going to be huge when dealing with clients as you're going to be able to give really custom solutions for whatever the requesting but knowing this let's go ahead and jump back to scheduler um and

[12:34] essentially just walk through a trigger we're going to say every day we want this automation to run so we're going to do every day trigger on weekends yes what time of day choose a time of day uh the time of day is going to be set up on your settings of the time zone so what do you have there so for us we're going to do noon going hit continue here and let's go ahe and test this trigger perfect so when you test a trigger and it comes back green like that that means there was no in running the process here anytime you test a trigger you're going to get a data record of whatever is

[13:06] relevant so what I mean by that is essentially it's going to give you an example of what it would look like if this got triggered as you see here today is Saturday um right now it's 5:00 p.m. where I'm at right now and it just gives you example of what the data would look like so in this context it wouldn't actually run because of the fact that I set it to run at noon so just to make that clearer and you get like a really cool trick right here if I added a filter and because of the fact that this flow only runs at noon it essentially if

[13:37] I put this um if put the time here and I put you know exactly matches it has to match noon so we'll do 12 it's not going to be able to pass through this because of the fact that right now is five so you know this wouldn't be triggered until tomorrow at noon as you see here the zap would not have continued really cool delete that come back over here this will trigger regardless and here's a really really cool feature that zapier just added I think today or yesterday but now uh this could come into this could be very powerful you have the ability to actually modify a data record

[14:10] and maybe input something custom like I could put anything in here I could put uh pizza right you can actually modify the underlying data that you plan on pushing through a flow which is really cool as this is going to allow you to do a little bit more testing and debugging with a potential app that maybe you can't get the Redd available data for now that is a natively integrated app called scheduler let's go ahead and change this to um a third party app like Twitter and look at a live data record here so we're going to do an event here we're going to say um new follower we do

[14:43] new follower of me and then we have our account integrated here we'll show you how to do that later in this tutorial we going toit continue here we going to test this trigger and then let's see who was followed me recently and as you see here we'll get multiple data points so I click this click this this click this we have all the readily available information for the underlying data point so what's great about this is that I know when I get a new follower on our Twitter account I'll be able to access their name their screen name their location their description I mean

[15:13] a ton of different information about that underlying data point that then we can use for the automation or integrate AI with okay so you got everything you need to know about the trigger let's start talking about actions and what better action to talk about other than chat gbt so we do chat gbt here when continue we're going to go and do conversation you're going to have your chat gbt account integrated here and what we're going to do is in our open AI tutorial we'll show you how to integrate the open AI block and the Chad gbt block so don't

[15:45] worry about that we'll have that all set up in the next lesson of this course but from here we have a ton of different stuff we can do with a chat gbt block so let's go ahead and walk through step by step everything you need to know about when it comes to a chat gbt action if you're familiar with our videos or you've seen our content before then you already know probably one of the most important parts of any Chad gbt block is going to be the user message this is where we're going to do our prompt engineering now one thing I want to point out before continuing this is there is a difference between regular

[16:16] chat gbt that you have like the chat bot with and the API calls that we're doing within zapier's platform the major difference is the fact of privacy so in the context of using the API like we're doing here open AI has stated in their records that essentially they don't look at the information that's being communicated here due to the fact that some of the information may be sensitive because we're dealing with data from other platforms now that being said if you use Chad gbt the traditional way they do have access to all those chats and they

[16:47] are grabbing that data and using it for their platform so keep that in mind also side note that could be a really good selling point to a business a business might be worried that hey you know AI looks great we want to use open AI we're worried because we don't want to give up our data within our backend because of the fact that open AI will use it if you make it clear to them that well in the context that we're using the API you don't actually have to worry about that and it is secured now we'll jump into prompt engineering and how to structure prompts in our next lesson here but from

[17:18] here let's go ahead and do the other variables here so we have username this isn't too important here you don't have to worry too much about username assistant name this isn't really important as well this is more of like when the output incurs what the individual or the chat bot is called uh in its output now the next variable point is assistant instructions the way we kind of look at this is like the cherry on top of the Sunday it's not fundamental but it still can really help with your prompt outputs so what you can

[17:48] do here for context is let's say you have an email or you're building out a automation which the purpose is to respond to customer messages you could set the assistant instruction to you are a customer representative of XYZ you know company now we get into the next most important part about the chat gbg block which is going to be the model and the differences between 3.5 and four think of 3.5 as more of grunt work this is going to be the one that you're going to use for more of data restructuring or finding data within a block of text so

[18:20] for example you can do a formatting where essentially you want the major points found within an article or an email or so on you could also use use a 3.5 Block in the context of finding sentiment analysis is this message positive is this message negative furthermore we can use a 3.5 block to find a piece of data within a text so find the email find the phone number where was this individual located very very simple and more specific fine-tune task should be associated with 3.5 and more comprehension task should be

[18:51] associated with four now you might be saying what do you mean by comprehension in this context I mean when you use four you are wanting to use for like the end output is going to be read by a human EG social media caption email response article Generation stuff that requires more comprehension and you know it's going to be read at at the End by a human rather than just grabbing a data point using it later in a flow like you would with 3.5 you're going to want to use four for now your next question might be okay I understand the

[19:22] difference between 3.5 and 4 what are all these options so what are all these other models these these other models are timestamp versions of the models so for example this model right here gbt 3.5 turbo 0613 that is the model from June 13th don't use it because of the fact that when they update it you're going to lose that model and that model is not going to exist anymore and this one for example is from March 1st come over here this is from June 13th and so on now the

[19:53] next question you might be asking yourself is okay I understand that that's from March 14th that's from June 13 what is the 16k so let's say you're using gbt 3.5 turbo but you're inputting a lot of text uh let's say you're inputting you know 10 pages of Google doc text and you're using 3.5 turbo and it's not really giving you effective answers this is where you're going to up it up to 3.5 turbo 16k as 16k identifies that it can handle more data inputs now to be clear with you as models keep

[20:25] increasing and getting better we're going to soon see a gbt 3.5 turbo 32k that's going to be able to handle much more data but for now to handle a lot of data in the context of AI and zapier you're going to want to use a gbt 3.5 turbo 16k now the third most important thing about a Chad gbg block is going to be your memory key this is going to ensure consistent outputs and consistent variety within your outputs now what do I mean by that traditionally when you're working with chat gbt and you're just chatting with it and it's referencing

[20:56] context that was said before in the chat think think of that as the memory key here you're basically creating a new chat with chat gbt but what if it's API here's why that's important when you are structuring outputs with chat gbt and you get one you really like maybe you get a solid email body based off the contextual data you provided but you don't have a memory key that means that you're going to be playing Russian roulette every single time you run that flow again because it doesn't know the consistent constraints that you want with its output if I add a memory key

[21:27] that means that that type of par paragraph that was outputed before would be outputed again but just with the different variables what do I mean by that essentially let's say you don't use a memory key and you get a really effective body and you say okay that's good I'm going to move on in theory next time you run that flow there is a possibility that emojis could be added there's a possibility that quotation marks could be added there's a possibility that the body of that paragraph three sentences out of the six are really bad and not what you wanted if you don't have a memory key this can

[21:58] encourage if you had a memory key and you like that first output you get consistent outputs like that following this you can get variety this is huge when it comes to content creation Now you can set it up where essentially you let's say you're creating social media captions you can add a parameter block that essentially says don't use previously said topics and it won't do it this is crazy because now essentially because it has a memory key is able to understand hey we've already said this in the past we don't want to say this again let's go and find a new topic or a

[22:29] new way to say this caption okay now moving on to the next input on a chat gbt block which is going to be your tokens tokens in this context is going to be how many you're willing to expend for a chat gbt output tokens is how open AI charges you based off usage so 1,000 tokens right now for 3.5 is like 0.002 cents I believe th000 tokens for gbt 4 is around 06 knowing this there's around a 2,900 difference in price between 3.5 and four

[23:02] therefore it is very important for you not only to continue to use 3.5 but use four in the correct context now the last two inputs we have here are temperature and top P these aren't too relevant when it comes to a automation all you really need to know about these is that if you lower temperature in top P you're going to get more consistent but less creative answers but if you heighten top PE and temperature you're going to get more creative answers but but some of those answers might be a little bit out of line of what you're looking for now that you understand the fundamentals between

[23:33] a chat gbt block on zapier there is another open AI block you're going to want to use and it's called open AI so they went ahead and separated the two and you're going to want to use Chad gbt AI with a chat gbt block and then all other functionalities of open AI with an open AI block essentially this is what I mean by that as you'll see here we have a ton of different options when it comes to open Ai and the abil ability for an action we have the ability to do app extensions check moderation create transcripts create translations generate

[24:05] image send prompt API request search embeddings so open AI use it more in the context of wanting to use whisper API or diay API and just for your reference whisper API is what's allows for translation and transcription and Del is for AI image generation so for example if I wanted to generate an image I would just do generate image here I'll hit continue choose my account hit continue here again and you know this is where I would input my my delay prompt here if you're

[24:36] familiar with the other video run like a dog uh on a beach we did that in one of the original videos we've done here at web Cafe AI um this allows you to choose how many image variations typically in the context of a a automation you would choose one and then obviously you would want to up it to the highest quality as possible hit continue Test action and we would be able to get an image uh let's go ahead and just do that so Test action here it's going provide us with a link we can open a new tab and just see what it looks like all right we can go ahead and grab this and as you see we got our AI generated image obviously I think as

[25:07] time goes on we're going to get even higher quality images it's going to get really cool when it comes to open Ai and the possibility to create content with it but for now this is how you essentially grab that image from the underline block and because I showed you how to do that of do I might as well show you what the output would look like for a chat gbt prompt here we're just going to say generate a social media caption for a dog company called walk walking and we're going to go ahead and continue that is not structured correct

[25:39] at all so don't worry we'll get into more prompt structuring in the next tutorial here but for now let's go ahead and see what our output would look like so as you see here this is what an output would look like it's going to tell you the model this is where the role was coming in earlier I was talking about it's not too important tells you the content and then you see here this is the output right here and like I said before as there'll be quotation marks you know obviously using a format of block here would be able to remove them uh from here you can come down here which is probably the more important part about the output which is going to be the tokens used and then the tokens remaining you don't have to worry about

[26:10] tokens remaining rarely do you ever exceed that for tokens used though that's going to be a great use case for understanding cost when you're finally done with a Automation and you want to run it on an Excel sheet but for now as you see here this phrase and this input and output cost me around 27 gbt 3.5 tokens so now you have an idea of a zap your flow you have an idea of a trigger an action and so on let's go ahead and jump into other parts of this platform the next most relevant part about this platform for an a automation agency is

[26:40] going to be your app section so this is where you're going to be able to connect third party apps that are prevalent with your flows so for example as you'll see in our next lesson here we're going to connect chat gbt we're going to connect open Ai and essentially when you're starting off at a automation agency you want to connect connect everything you can think that you'll use in that agency so a good one obviously would just be putting out Google here connecting your ads your analytics your drive your forms your sheets everything associated with Google you'll probably use with AI

[27:10] automation or you'll use it to test with it so it's always a good idea to connect all of the apps that you plan on using with your agency onto zapier so the next major tab you'll find in your zapier account is going to be zap history this is going to be like an admin view of all the task run all the task audio and so on as you see here we have the task usage and then we come over here there are some zaps that ran today if I go over here to zap runs it's going to show me all the zaps run today whether there was a zap that was run but it was filtered um and then if there was a

[27:42] success or not on top of that it's going to tell me how many tasks were burned for that specific run so as you see here we got a three task eight task one task and so on you can also set it up so you can search by zap you can search by apps you can search by folders so maybe you have a specific client they're doing a lot of usage now you can go ahead and input that client's name here and you'll be able to get the amount of Zap task and zap usage that you've had with that client for x amount of time now one cool thing I want to

[28:13] show you about zap usage and the ability to make it so it's even more automated and less manual input is if you come over to history you will see a option for auto replay this is really important when you're dealing with a lot of zaps because of the the fact that not only are we dealing with AI we're dealing with just zapier as a platform sometimes you'll get errors that aren't even relevant and what do I mean by relevant is there's nothing you could have done the actual automation or a automation is correct and everything you inputed

[28:44] should work sometimes there is just an error due to maybe chat gbt running out of time there might have been too much usage on the server maybe zap you're an issue in they're backend and you get an air therefore causing the automation to stop prematurely but if you keep auto replay on this is going to allow it so essentially that the zap that may have faed as an air For No Good Reason will automatically rerun and run correctly this time now the next tab here is going to be the explore tab this is going to allow you to essentially look through uh

[29:16] different examples of automations that have been built by other people um I will say a lot of these are very basic and surface level they don't really get that complicated at all but it's always a good idea to check it out cuz maybe you'll get an idea from a very basic automation you see here this we have a GI help section and then the current plan that your business is currently on in the bottom left this allows you to track how many tasks you've used up to this point now another section that can be found on zapier is tables this is going to give you the ability to

[29:47] essentially create Excel sheets that have the ability to have ai automation or automation General better integrated into it the way you create a table is simply go to the tables tab create a name for it so we can put test here here we could import data that maybe we already have with a CSV air table or Google sheet and then we can choose a template right now the only template available for us is a to-do list now in this current context the best use for a table is for more of lead management therefore when maybe an individual inputs on a Google form or sends an

[30:17] email it will automatically fill the tabs here with the relevant information and stuff like this right here gives you the ability to basically set off an automation with that Associated data for now all you really need to know about tables is it's going to be a very powerful tool in the future once we have more capabilities when it comes to data uploading and the ability to really manipulate what's going on here now interfaces is a great feature for creating landing pages or chat Bots for

[30:48] potential clients as you see here there is templates that are already provided here such as a new client intake a lead capture a request portal and so on I see interfaces as a way to maybe expand the services you could offer to a potential business due to the fact that it seems like zapier is beginning to bolster the abilities of this platform as it's still in its early stages but there is still a lot of value you can grab here okay so now let's understand when you look at a zap run what all this information means right at the bat you'll see here the

[31:18] amount of task expended for usage but as a rule of thumb all you really need to know when it comes to task usage is that if there is four blocks including the trigger as you see here you just have to minus one so the fact that there's four blocks here minus one three tax will be expended due to the fact that we have one to three task in total the next most important part about looking at a zap's run is you can look at data in and out this can help with debugging uh for example we can see here the data in here

[31:48] was uh this right here and then the data out went ahead and restructured it and split it between the caption and the hashtags you can see that with the DBT as well with the data in being the original user message and then the data out being what the user message um response was okay we went ahead and covered everything you need to know about the zapier platform and getting started with it now you have an ability to really build out a automation flows get a handle on this platform we plan on jumping into this next lesson is

[32:19] everything you need to know about open AI we're going to integrate Chad gbt and open AI with zappier so we can start building out these flows but without further Ado I'll see you in the next lesson