here's how you build anything with the brand new make.com AI agents i have spent hours going deep on these AI agents so you don't have to and I've helped thousands of business owners now absolutely crush it with AI automations and in this video I'm going to show you a no fluff sevenstep process to build reliable powerful make AI agents and then I'm going to show you how that compares to NA10 and when you should use which agent for which and at the end of the video I'm going to show you what Make.com team have personally told me about these AI agents and also some of the top performing use cases for AI agents to transform your business or things that you can sell to other businesses so if you haven't already grab that coffee and let's dive straight in and I kind of want to just start off with guys about the pricing so how does the pricing work with make.com AI agents versus NHM well the pricing is fairly standard that as you know and it's self-hosted but for make it has no change to the actual cost that you pay so whatever you're currently paying for your operations and processes will remain the same because with make.com agents effectively what you do is you give it an API key to your language model which means that you're the one that covers the costs for when it runs things and in that way it runs exactly the same way as NAM beautiful so let's start off guys I guess with actually building an agent step by step and then I'm going to cover my specific thoughts on when you should use it so to do that let's begin with the beautiful seven steps and of course we're in Simpsons Land we're having a great time now before we do this I wouldn't build and make AI agent until you do the following and the first thing we need to determine here guys is whether we need an agent or an automation it's a very important distinction so a couple of questions I've put down here and again this whole PDF will be down below for you in the description just click a button and you can access it immediately um so first one is does it need complex decision- making so when you're thinking about okay I want to build something ask yourself these questions so if it's workflows that require autonomous decision trees or conditional logic based on different inputs you may need an AI agent versus just something that goes left to right because remember in make.com AI a basically all automations are linear and they go left to right but if you can define what that is and it isn't complex logic you may be better off just building a simple automation second of all does it coordinate multiple steps intelligently so your automation needs to understand context across steps and adapt accordingly does it need to have and bundle multiple things together you need to think about does it need to select and use different tools so for example in this one here I've got a make AI agent and we're going to build one together of course but let's just say for example you have a Gmail module a calendar module a telegram a research module does it need to make intelligent decisions about where it wants to go based on the input so if I have a personal assistant and I say hey go forth and uh oh check my emails and oh brokerage you do X that is multiple different routes now you can hardcode that with with basically uh logic and and systems but actually that's the kind of flow that would prefer an agent over not so if it's got multi multi steps that's important fourthly does it maintain conversational context are you having a chat with it and does it need to have memory if so use agents and I'm also going to tell you as well guys about me make.com memory because they've basically solved it and massively simplified it if you compare it to any for instance when you want to retrieve information actually in fact if you click on AI agent hero you see you got this memory you can choose multiple things in make it's just there automatically it just kind of works on autopilot and it retain information i did ask the make.com team about how that works what's the limits what's capacity they haven't got back to me but as soon as they do I'll put some information under this video so you can check it but you don't need to worry about memory remember their direction of travel like they told in a podcast is basically to make it as simple as humanly possible beautiful and finally does it need to work across different platforms coordinate multiple various action systems and intelligent evaluation so just think about first of all is the agent is automation most of the times it's going to be automation but when we need an agent we proceed down to the beautiful step two now step two is to create the AI agent cool so how do we actually freaking do that well what we're going to do is we are going to head on over to this scenario so we click this button and open it up and it will load for beautifully and guys what I'm going to do in this video is build a personal assistant agent a very simple one because I just want to explain the technology now all you do is come over and click on create agent like so and we're going to call this one here we're going to call this Jack virtual assistant beautiful and of course we'll give it an emoji a rocket now I'm going to pick the model so we're going to use GPT 4.1 which came out yesterday and is blowing everybody's mind so I want you to come up here on add click on connection type and we are going to select openai beautiful if you haven't created a connection yet very simply how you do that is the following you're going to head over to platform.openai whoops we're not on uh this guy platform.openai.com and then you're going to come down to the bottom left and you're going to click on API keys and then simply create a new key you'll have a load of information create secret key it'll pop up you're going to copy that data there and then you're going to come back over to your make to come scenario and add in your API key right there which is exactly what we're going to do let's copy this guy come over and pop it in there beautiful and then you click on save don't worry about organization ID now what we want to do to use GPT 4.1 it hasn't appeared yet as you can see right so to do that we click on map and then you just enter in this code GPT 4.1 and then what we need to do next let's go back to our beautiful guide to figure out what the next step is so we have headed over there beautiful we have chosen a model we've gone for 4.1 uh and then give it a system prompt cool so to do this guys I just recommend that you actually ask ChatGpt to help you with that so let's head over there now beautiful chat GPT okay sweet so I'm going to say hey create me a uh AI agent prompt a simple AI agent prompt and I'm going to say hey uh input will be a Slack me we'll call it maybe a Telegram message okay and then we say tools and then let's just give it two tools let's just say for example ability to search uh the internet and then secondly uh we'll also say its ability to do something really interesting like I don't know uh create a draft email as a quick for instance okay and of course you want to say this is a personal VA beautiful let this rock and roll around the background oh you know we should have just used 4.1 right where's 4.1 oh my 4.5 4 oh my goodness right we come down cool so we've got this personal agent prompt which is sick so what we're going to do guys very simple is copy this you can go really into detail with it and I' I've got things on that but just to keep this one beautifully short and simple I'm going to come down here um all right sweet so let's come down here copy this guy beautiful head back over then to our AI agent and we enter in the system prompt the system prompt is kind of like it's hardwiring it's like core functions obviously Meg's going to write things behind this as well but effectively it's what we're doing okay then we click on save so you have effective we created your agent now the first thing I want you to understand about is some differences okay so look if we look at make AI agents how do they actually differ than NA10 and if I come over here to show you these things with make agents and how they differ the here's a cool thing I want you to understand conceptually about it nam is nonlinear in other words we have an agent here okay and he has access to multiple tools and we plug in those multiple tools in a kind of spiderweb and that agent reaches over and does this and you can see it and you can add things in in a nonlinear way okay but by contrast if you look at make agents how they work effectively you create a master scenario so we've just created our AI agent okay that's this guy right here then effectively what you do for its tools is you create multiple other linear scenarios in make.com okay this one could be the ability to search internet this one here could be its ability to create emails this one could be its ability to create calendars so you're kind of think of them as weapons like that's its mech suit that's its gun that's its X that's its Y and effectively once you build these this library that's a big thing about this these AI agents you're building a library a suite of tools and then once you've built that hard work and investment then any agent you build you can just reach over and grab the tools so the key difference here with make.com versus NA10 is the fact that at make you need to build up your repository your your war chest of tools once you've built those up you can unplug and play them in any way that you want to na10 you build them on the fly with the agent as you go every time so that's how this needs to work we need to first of all build the agent scenario and then we need to build lots of mini on demand scenarios i personally wouldn't have called them scenarios i would kind of separate them out a little bit in tools and I would like to be able to do that in the actual uh automation itself but again I'll explain more about that in detail but that's just to give you some conceptual framework now look if I come back over to our steps we need to give it system prompt and we've done that so now we're on to step three of seven which is to choose a trigger now ultimately guys um this is always great practice when you're building agents or any scenario at all is to think about the trigger and I've detailed some example triggers down here telegram uh web hooks for example the web hook could be something from lovable we could create a beautiful gorgeous app let's say we want to chat with videos and then we want to call an agent in make.com that does several other beautiful things one I'm going to go with on this guys is because we're going to build a virtual assistant i'm going to go with a Telegram hook and we're going to set that up together and crush it so how do we flip and do that well I'll cover that in a second we do that so we're going to pick the trigger which in this case is going to be Telegram obviously you can do different things like payment received just have a look at these and you'll see maybe you'll get some different ideas about the kinds of things you can do with agents but fundamentally start with a trigger and we also want a defined output cool all right sweet so now what's the next thing we need to do we need to create a scenario with an agent so let's we're going to create a scenario we're going to add in the trigger and the agent and then we're going to connect the inputs all right that sounds pretty cool let's go over to our new scenario and we are going to call this one here let's call this one AI agent in fact Jack's AI agent VA okay because who doesn't need a Maverick right who we want this this future of Iron Man with something that does everything possible for us is flipping epic all right cool so what do we do we need to add in a trigger so for us we are going to go with Telegram telegram bot watch updates i'd recommend Telegram by the way for automations because it is the API is freaking amazing it's so configurable you can send images and audios and loads of stuff whatsapp is very defensive like I'm getting my guard up like I'm in an MMA fight right now but it's very defensive so to do that we're going to create we're going to head over to Telegram you're going to go to Telegram.botfather um the botfather i love the kind of energy there click on start bot we can open Telegram that's cool and then you're going to come down here to start and then you're going to go with for slashcreate i think it's new bot I think isn't it new bot beautiful okay all right it's going to do some magic for us in the background we're going to call it we're going to call this one Jack VA with loads of A's uh bot obviously must end in bot call it whatever you want to um yeah we did end it in bot right okay it's got to be uh with a cap okay there we go that's already taken un Okay we're just going to give it a random freaking name here cool gives us an access token copy that access token then come back over to your beautiful scenario in make which is right over here you're going to add for web hook you're going to click on add again and add in a token and we'll call this one Jack um VA bot okay beautiful then click on save and that should be phenomenal and that will work for us beautifully great so now effectively what's going to happen is the second that we get a message uh this scenario is going to trigger for us so if I come back over to our phenomenal Telegram here and I want to have a chat with it I should be able to uh click a button here cool and I say "Hey man hey man." All right we come over uh we want to uh run this one module's not set up click on save run this guy here okay let's send a new message again hey cool guys and as you can see we've got this text hey which is flipping awesome so let's go back to our seven steps just to make sure we're crossing off so what have we done we've created a scenario beautiful we need to add in the trigger and the agent okay we're adding in the trigger now guys let's add in our cool agent so type in AI agent and make AI agents run an agent beautiful okay cool right it's rocking up now we need to choose our agent and we are looking for Jack virtual assistant very cool indeed okay a couple things just to understand here guys um thread ID is going to be the chat that we had on the left hand side so effectively what this will enable it to do this is what I mean by memory so what we're saying to this AI agent is look I'm going to be chatting to you for a little while all right I'm going to give you a unique ID that tells you it's this chat so you can kind of retain it so I ask you a what did I just ask you it will know and it will retain context this is what I mean by solving memory so we're going to come down and find the thread ID for us real quick and I believe that is just simply here business message chat ID cool then additional system instructions here guys um is basically anything as we're running the scenario we're thinking you know I need you to kind of do this think about you've got your basic prompt that we created in the agent screen if you run this let's say agent in multiple scenarios you can add a little flavors and feels to them but if you find that say hey I want it in HTML versus what you gave me just add that as additional convenient way to add new things cool and now we've got the message which is going to be the input so in other words we need to connect it to the input so we click on message one and what are we looking for is simply the text here beautiful that's it so we've got the message text that's freaking phenomenal we like that continue to run store on agent is working no well you'll rarely need that but I think for for just to keep it simple we're going to put on um message text for them then click on save phenomenal so we come back over to our scenario here we've added a trigger and we've connected the inputs beautiful now on to step five of seven that is to create tools of course one thing that we must do as well just before we move forward is name this which we have here Jax AI agent V8 beautiful now we need to develop scenarios for our use case each scenario is a tool so we're going to create two scenarios here okay we are going to create uh let's just bring in this guy over here one is going to be the ability to search the internet so I'm call this one here internet search very simple right and what do we do with internet search well first of all we're going to click on this plus button here we are going to go for perplexity AI and we're going to click on create a chat completion and I want to do maybe we do deep research with this so you got Sonar Deep Reasoning Pro let's say that I I want full beautiful information okay now we've got messages max tokens pressure so one of the key things that we need to do here guys is set custom inputs and custom outputs so you'll notice this down here which is scenario inputs and outputs so don't worry about this don't get too confused effectively what we're doing here is we're able because you know classically when you build a module you you you enter things like this is what this module said this is what that module said here we're going to create our own unique inputs and that is that so we can use it with this beautiful AI agent because he is going to reach over into this scenario and use those inputs so for example input structure here is I'm going to say uh the name which is going to be the name of the input we're going to say basically just simply research this that's it description uh basically search the internet for this information that's basically it so in other words what is what is it and what do you want it to do beautiful uh and then we're going to say yes that's required and then we're going to do scenario outputs which is simply going to be research output beautiful the results [Music] from the internet search really cool right phenomenal and then guys click on save so there's our scenar now when I click on Plexi AI and I come down to messages what I will be able to do in content is come over here to this section here and let's see if we can find it i think it's a curly brackets guys okay we don't see it yet that's interesting right so I wonder if we just very simply need to uh first of all change the trigger so right click on this uh actually click on this guy instead of at regular intervals come down and click on demand right that means that this is now we need to activate the scenario let me just save this guy first that's okay i'm going to uh clone the module i'm going to delete this guy and the reason I'm doing that is sometimes now we've set the custom variables um maybe they don't appear sometimes i found that's just a slight little quirk it has so come down to messages look at content and then again come to the curly brackets and look at this guys scenario inputs research this has appeared so effectively this is how we connect it to the agent okay now the role is just going to be simple user and guys that is literally as simple as it is now we need to click on one of the module which is going to be to return output in other words hey bro this is what I want you to send back so we click on this and what we're doing again just clicking curly brackets and it's not just our input in fact we want to come back over here come to choices come down to message come down to content and then click save guys and that is literally it and is this live needs to be live that's beautiful we click this so it's on demand and we click save and we're just going to refresh the page just to make sure that it's all working beautifully for us give that a hot second likes to keep us in suspense a little bit sometimes okay cool that's done so now I've done that let me just check how we got we have developed scenario for tools beautiful we have named the scenario something relevant incredible we've assigned inputs and outputs for scenario amazing and we've set it to run on demand okay nice now we need to add our tools to the agent so I'm going to do one other tool very quick so this is our core agent this is one of our sub guys one of his tools and let's add one more so for this one let's just do something like Gmail uh maybe we could do something like um let's have a look to see what we can do maybe go down to create a draft as a quick for instance now what we want to do again is come down here to scenario inputs and add input structure so let's just say something like send uh a recipient who we are sending the email to beautiful and then we want to Yes that's required and then we want to say something like subject right and say the subject of the email and then what else would be cool probably body right uh body so you say hey the body of the email all right then click save and then we want to click on this and go to scenario outputs and then maybe in the outputs we could probably say something like confirmation row that the email has been sent okay beautiful come down and yes that it's required then we click on save and that is beautiful cool so click on create a draft uh we need to verify connection i'll just do that very quickly beautiful so I've confirmed that of course we don't need to click off it and click back on and I'm just going to say hey dude just save this in a draft folder for me beautiful so let's come down to drafts excellent and now again we just enter in all these custom things so what is the email address so come over here this is recipient beautiful subject is come over to curly brackets subject you might have guessed it and then content is come over to curly brackets the body okay awesome and then we click on save and that is freaking it guys then we come over here of course and click on return output which is freaking awesome and this is confirmation confirmation email has been sent maybe if we could run this as a quick uh example so run this module only real quick okay so that's interesting you can't run it as an example so confirmation i'm just say hey the email uh has been sent here's the message ID okay cool it's so incredible how the AI agent space is heating up so much right cool and then we want to obviously right click this guy now we want to make sure that Oops click on the clock click on regular intervals click on demand click on save click on activate scenario and we'll say here create draft gmail draft and the cool thing is once you've created these tool systems and configured them you can pretty much do whatever you want to which is nice and I would turn this one here snar is not saved yet okay we'll save it man don't you worry about it man we'll save you and click on demand cool and guys I'm also going to get the uh the guy that created AI agents from make.com on the podcast if you got any questions down below let me know and I'll make sure that we answer it and have a good conversation cool so we've done that we've created our two tools now right we've created internet search and we create the build to create emails so what do we do next now we want to add these tools to our agent.un okay cool so how do we do that now we go back over to our virtual assistant now we see the system tools down here actually do you know what you can also do this uh in the scenario itself i believe I believe you can look additional tools click on add tools select your scenario uh so then we can just search for the things that we had right so this one is internet search so come down here internet search and it'll say active so you must only add internet search searches internet beautiful then we add in another tool and then we come down here and this one of course was Gmail I think create Gmail draft uh basically creates a draft in Gmail cool all right that sounds pretty decent and then click on save i think that's good oh actually one thing I will just say actually is additional system instructions is format your outputs uh for Telegram messages uh use emojis because you know I'm a sucker for emojis guys you know what I'm like for design i I spend way I would say I spent way too long on design but I don't i just think it's really important use emojis and summarize what you've done call me the Night Rider my phone by the way is goes by Night Rider which is funny people always joke about but dude like you know I used to watch the the freaking thing in like the ' 90s and early 90s this is my VA my VA is a Night Rider i like that then we click on save now what we want to finally do is something to send us back the updates because we need to know what's going on right so we go to Telegram okay and then we want to click on I believe where is it now it's going to be edit uh where is it now send a text message or reply then you need to enter a chat ID guys um the way that you find that chat ID I'll just find that for you real quick you're going to go in chat guys new message raw data bot and all you're going to do is forward slashstart and then it will give you a code which is your chat ID then when you come over guys just enter in this chat ID section here and then for text we simply want the output from this which is the response and then click save and you're ready to go beautiful so let's actually freaking test if this works all right so I'm going to click run once obviously what we would do is um wait for this so let's give it a question okay where it's going to use one of the agents so if I come over to our Jackbot and I'm going to say some cool stuff to it i'm going to say "Hey uh what are the top five coffee shops in leads?" Bam i'm going to hit enter now look at this make AI agents running in the background this is pretty freaking cool now what you'll see let's see if it's using any of our scenarios right so what we will see and you won't see it on this particular screen but if I uh come back out this is what we never tell you actually but look if I refresh this scenario check this out and refresh this guy oh hang on i need to save this first you can actually see them being used um you'll see it here once if it's active and being used uh let's combat this guy okay beautiful so make AI agent is working in the background sometimes you can actually see okay no execution is currently running that's interesting so I wonder if in this instance it's just trying to figure out itself and will send us a message back guys what I notic is I didn't set this scenario to run on so I'm just going to come back and I'm going to redo this scenario so that is completely my fault i should ask as well just to make sure that both scenarios are on so let me come back to our seven steps and let me just say okay we've connected our tools to the agents all right and I'm going to say here ensure all scenarios are active all right and now we've done that cool so the guys this list is getting better every single second and I've got some really cool uh data to show down here too all right cool so let's come back over then to the scenario let's then come back and run the scenario once let's wait for new data let's go over to Telegram i'm going to say ask it the same questions um I might say as of 2025 cool okay we hit enter bam this kicks off make AI starts doing its searches and then we should be able to see if an active scenario starts working okay no execution is currently working yet no execution is currently on this one let's come back over here and see what it's doing and actually guys before I do that I'm just going to add in the tools again at the core AI agent page to make sure it's got the full access because I'm not seeing it in the system tools here so let me just add those in again we'll just do uh Gmail which is active beautiful uh and we'll add this guy in uh basically create a draft in Gmail because one of the reasons I just want to do this before we run the scenario is the fact that Okay so add that in okay sweet and then add the other one which is going to be uh search internet search search is internet right beautiful okay fantastic and I'm going to click save on this guy so we're all ready to rock and roll beautiful right let's come back over and let's try this guy i'm going to hit run once and we're going to give it the beautiful scenario questions let's come down and give it the question Telegram real quick uh oops as of 2025 question mark bam okay send that off make a agent start running in the background and then we should be able to see when these active scenarios start pipping and popping over here all right guys and just like that here's a top private coffee response for Night Rider Lane Espresso Lee Monise Mrs athes mrs athes has shut down recently that it's not on the internet yet but that was a real freaking shame cuz they were awesome tundre coffee and coffee boy beautiful that's really sweet and did we get the message on Telegram we did it because of what unauthorized oh I need to add in the API connection let me do that very quickly going to come back over to the bot father get my access token and we need to add it in beautiful okay guys now we're actually working and things are making beautiful sense so what do we want to do next i guess let's ask it now to create a draft email for us so click on run once real quick uh wait for the data and then why don't I click on telegram and I could say something like hey uh create me a draft email to me at Magoo Enterprises uh.com it's about uh how great the coffee was at Mrs athers uh ask him if he can grab me a matcha latte uh beautiful and he'll click on enter bam that's gone off now oops i've sent that to the bot father he ain't going to really understand what that's about right okay we come back over let's go over to our WhatsApp button and ask this guy beautiful we sent it over this is our triggers make AI agents now running in the background now guys one interesting thing about I'm finding with the make AI agency is some people reporting that it's more effective which is very cool we like that um and we got it back already so let's come back over and see what it said beautiful draft email for boom is ready complaint after coffee asked if you got actual latte let me know if you want to tweak it night rider okay well let's get over to Gmail and check out the results I guess and guys would you look at this meoo Enterprises hey Magoo just want to say the coffee is fantastic you're right about it all right guys so now we've got it working across two different functions we've got it here who can do and essentially what we do is just stack in all the things and then we want to reduce the operation maximums so come over to this page which is Oops so we're going to come over to this page and we kind of look at this agent settings thing here so we've got the model GP 4.1 which is awesome max output tokens this is the maximum amount of stuff that you would want to output in terms of tokens recursion limit and then iterations from history count so the recursion limit here guys is effectively think of it as like the number of times it can call itself the more complex the scenario the higher the limit the simpler it is the lower it should be so I'm going to lower this way down to like 80 something like that and the iterations from history count is kind of like how many past conversations should it think about um when it gives you the response there which is cool so I think we're happy with 10 10 is fine again if it's like a chatbased model you want to increase it up uh significantly and then come down and of course we've reduced the maximum number of operations now guys I want to give you uh two uh two bonus things right now number one is use cases AI agents and the second one is going to be my general thoughts on this and when you should use each so okay basically this is in beta obviously it's not perfect yet its memory works for me sometimes not all the time so when I've run it sometimes it can remember exactly what I said and other times it doesn't it's also weird in the sense that it seems to be able to remember things that have happened like months ago that I haven't even used in an AI agent like I asked it um what did I ask you about this email and one of the things it told me was a message I sent to Adam months ago so I don't know if this is just what the hell that's about but like if it's I don't understand the data so when we get them on the podcast I'm going to say "Hey dude how's this all flowing up?" So make all the right decisions all the right moves you have to remember this is in beta right now functionally it can do everything that N8N can do in that sense right so if you actually wanted to build it in make.com the question is which one do you want to use you can do it's cool some people saying they finding the operations a bit crisper which is cool but actually it can do the same sorts of things um I'd say that to be honest building NA10 is probably easier just because I don't need to go from like scenario to scenario to scenario right i can kind of just do it in one so one of the things I would love make to do is kind of let us build the tools in the environment that we're in like don't make us click from one chapter to another i want to be able to do it all in one place the other thing my other feedback would be in NAN when you're running an agent you can actually see what the agent is doing and you can see oh it used Gmail oh it did this i like that visibility just like an Uber I can see where my guy is i can see dude he's there in two minutes he's over there he's gone to the wrong location again you can do all that sort of stuff and make.com that's kind of blind at the moment so you're kind of just trusting that it's doing it i don't get to see the inner workings of the agent i can't look behind the lens and say "What's he doing what coffee is he drinking?" Flat white if you're wondering so it's like I'd like to see those additional details in there um so those are the things I want to see i think we're going to see how quick they update and upgrade these agents do they work yes can it do the same stuff yes but you're not going to get the full level of transparency in terms of what's happening behind it when it does complete a scenario you can drill down into it and click buttons to see all the stuff that it's doing uh I wonder if I can pull up real quick okay so if you come down you start clicking into execution steps you can see all the stuff that it does right instructions there we go look at this you can see everything it's doing right all these little execution steps it's doing i've just got to click lots of stuff like it's got my HTML tags in i've just got to click a lot of buttons so it's there i just have to do extra steps so on that basis I would say look both work both work absolutely fine make the other advantage you got with make is you can start to pull this into the grid the grid being that beautiful thing uh here where you can start to stack and zoom in and see how all your snares connect together so ultimately which one should you use costwise guys it is no difference on the cost at all because you're just using your own API credits na10 if you want more visibility what it's working at the NA10 AI agent is more mature at the moment to be completely fair with you it just I get to see what it's doing i get to see it's thinking these are two things I'd love to see make.com add to these agents um but actually if you're someone who's got your you love building in make you've got all your stuff in make guys you can use these today and absolutely crush it so based on that you can decide which of the two that you would absolutely like to use and then one final thing I want to add in there as well at the end this oh and also I I spoke to one of the bros at make and he said like look they're in the feedback stage they're not saying this is the holy grail go forth and use these forever it's like dude we put it into the market we want to see what you think this is what I think this is what I want to see i want more transparency i want to see the tools in the same scenario so I don't want to have to go to different scenarios to build this stuff i want it all in one i should be able to build my agent and have it all its tools in one place that would be freaking sick and I want to see more what it's doing because make the thing that's cool about Meg it's got these linear agents right but it's also like that is cool i do want to see you know I I want I want I think that's a good opportunity to like build on it and take it from good to great so beautiful first product i love it it works it's functional that's how you build it that's some stuff that I personally would love to see and I'll be asking Tim about when we uh when we chat to them cool now some things is I did a big survey for all the AI automation agencies in my community we've got thousands of them and I basically asked like 16 questions i did where you getting customers from what uh things we sell to people i'm going to do a full video on all the data I got from that but I'm going to add an SPF here for you um just databased AI agent use cases like the actual things people are paying for right now and businesses are using right now are here so stop and customer experience operations workflows sales marketing finance compliance product development human resources I'll put all that down below so you can check that out and have a rock and rolling good time if you haven't seen this video on screen about everything I learned uh building a sevenigure AI automation business I'd encourage you to go check that out but in any case I hope you found this video helpful have a beautiful week and I'll see you in the next