How to add AI agents into Slack (no code)

Rory Ridgers shows how to build a Slack app that runs Relevance AI agents without code. He creates a content generator using GPT-4o Mini, sets up OAuth and event subscriptions, then uses Make.com to connect Slack messages to the AI agent's API and post results back to the channel.
Chapters
A lot of you guys, if you run your own business, will use something called Slack. And if you've never used Slack before, what it is, it's pretty much a space that allows you to talk to your team, manage your upcoming projects and tasks and things like that. And you've probably heard before of a way of integrating AI into Slack. You can talk with AI agents inside of your Slack channels. And if you've watched this channel before, you know that I like to use Relevance AI. So in this video, I want to show you a really, really simple way that you can add Relevance AI agents into your Slack channels. You can either have a separate channel for these AI agents, or you can just them. So you can the name of the AI agent and then be able to talk to them. This can obviously be used for things like web scrapers or agents with custom knowledge bases on your business. So if you aren't too sure on something, if you're talking to a team member, you can just the bot, or you can go into the bot's channel and retrieve some of this knowledge on your business or something like that. Or it can be used for pretty much anything you can kind of think of that you can do with AI agents. You can integrate them into Slack. If you haven't seen me before, my name's Rory Ridgers and I'm an AI enthusiast and I'm passionate about helping business owners integrate AI into their business. So if you're looking to get ahead of the curve and integrate AI into your business, then do consider subscribing to my channel and following along for the video. So in this video, I'm keeping it nice and simple, and what I've done is I've just made a really, really simple Relevance AI agent that just generates content for things like blogs and social media posts. So yeah, just keeping it really, really simple, but I just want to give you guys an example of how you can add these Relevance AI agents into your Slack channels. And also something to mention is that you can do this with pre-existing agents that you have, so you have to remake agents or anything like that. If you have any agents on Relevance AI that you want to add into Slack, I can show you exactly how to do that in this video. It's a really, really simple setup. Okay, so in this video we're going to be using Slack, of course, Make.com, Relevance AI, and then also the Slack API. And if you haven't heard of some of them before, don't worry, I'll explain exactly how to use them when we get onto it. But to begin with, we're going to be in Relevance AI, and we need to make this content generator agent able to write these blogs and social media posts and things like that. So to do that, we just need one really, really simple tool, and it's this tool here, the content generator tool. And now this is a custom tool, but it's really simple to make. I'll show you inside of it. So as you can see, we've only got one user input, and that's just the input from the Slack. So whatever message they put into the Slack channel, that's going to be used as the input here. And then we've just got a simple LLM step using GPT-4o Mini. And you can see the prompt is, you're an expert on writing short and concise copy based off a given prompt. This copy is mainly used for websites and social media, so make it SEO-rich and compelling. And I've just said, um, never ask follow-up questions, never exceed 150 words, and use the user input as the prompt. So that's what's going to be generating the content in this case. And then this step here, the API step, this is what's going to connect it to Slack. So you can see we've just got a base API step at the moment. This is how it should look when you first put it down. But before we get on to that, I need to show you some steps in the Slack API and then also a little Make.com setup. So if we come over to the Slack API, you can see this is what you should be greeted with when you first come on to Slack. So what you're going to do is just going to hit your apps over here, and then I've already got one in here, but you shouldn't have any in here unless of course you use the Slack API before but yours should be blank. So what I'm going to do is I'm going to hit create new app, and then you should be met with this. And I'm going to click from scratch. The app name, we can just call it Content Generator. We should call it Content Generator. Pick a workspace, and then Autonomously AI, that's my workspace on Slack. So it makes the app. As you can see, this is what you're met with. You're going to come down to Zero Auth and Permissions, and then there's two redirect URLs that you're going to need to put in here. So this is the first one, and this is just going to link it to make.com. So we're going to hit Add and then add one more. And then this is the second one, we've literally just changed the 2 to a 3. It's already really simple. Just hit Add, hit Save URLs, and then what we've got to do is go over to Event Subscriptions. So just underneath And over here we're just going to turn this on. The request URL, so I'm going to show you what to put in there now. So I've come over to Make and I've just set up a new scenario. And if you haven't used Make before, it's just a really, really simple automation platform that connects with loads of different apps and allows you to have webhooks. So it's just custom URLs and they can be called upon from pretty much any sort of API or software. So what we're going to do, going to do a webhook, custom webhook. And then choose a hook. We can just do anything, we can say content generator, hit save, and then it's going to give you this URL here. And what we're going to need to do with that is we copy it, and then that's what we put in this request URL. So we paste that in, it's going to verify, and then we're going to get an error, which is normal, because what we have to do is come back over to Make, add another module here, webhooks, and we're going to go webhook response. Status 200, and the body, we're just going to put the challenge in there. Hit OK. And then what that's going to do is when the webhook's triggered, it's just going to return a response. That means it should be all good on the Slack API end. So we can just hit retry. Okay, so it's saying there's an error again, and yeah, it's because, it's because the scenario isn't turned on. So if we turn on, that means whenever the data arrives, it's now going to automatically trigger this scenario. Which is what we want. So one more time, if we hit retry, and you can see, yeah, verified. And then so once that's verified, what we're going to do is subscribe to bot events. And this is basically just putting in how we want the bot to be triggered. So in this case, we're going to do it so it has its own separate channel, but you can do it where you can just mention it. So you can say @ContentGeneratorAgent or whoever you choose to call it. But in this case, we're just going to make its own separate channel. So we click add bot user event. If you just type in channels, we're going to want to click on message channels and then save changes. And then if you just come over to install back up at the top here, you can hit install to my workspace, Autonomously AI, and it's just going to ask to allow some permissions. So we just hit allow, give that a second to load, and then yeah, success. That's all done. Now that means that's installed to our Slack channels. Now we can come over to Slack and we can make a new channel. So we're going to click add new channels, create new channel. Name. So for the name, I've just said Content Generator Agent. And then we're going to set it to public so everyone can see it, but obviously that's up to you. And add people, we're just going to skip that for now. And then now to set this agent up so it's connected with this channel, what we do, hit the three dots in the corner here, go to edit settings, and then you should see your integrations tab here. And then there's a little apps bit here, and we're just going to click add an app. And then sometimes these can take a little bit of time to load up, but this one's loaded pretty quick. You can see the Content Generator here, so we're just going to add that to our channel now. Yeah, as you can see, joined the Content Generator agent. Content Generator has joined too. So our agent has now joined the channel, but now we need to set up so that's actually talking to the Relevance AI agent. And we're just going to do that by making an agent inside of Relevance AI. So if you don't know how to do that, you just come over to the little Agents tab here and you'd click on New. But in this case, I've already got an agent here, a content writer, so I'm just going to hit Edit. And this is what you'll be met with. I've already gone ahead and done it just because this one's really, really simple. All it is for me is just the name, content writer, a little description. And now these descriptions do kind of matter because if you go on later to use multi-agent systems, the other agents are going to use this description to be able to determine what agent is which. So Do bear that in mind. These descriptions can be kind of important, but for now I've just done a simple one: write simple copy for websites and social media. And then the core instructions, really, really simple. All I've said is use the content generator tool provided, take the user's inputs and write content. That's kind of all it needs to know. There's no special prompting with this one. And for the tools, we've got that content generator tool that I showed you earlier, and we're going to come back to that. But for now, what we want to do is go back over to the agent profile, and we want to be able to connect this with the Slack API to be able to talk in the Slack channel. And we do this through Make.com via a webhook. So what we're going to do, we're not going to click on webhook, we click on API. You can see we get an endpoint, we get an API key that we can generate, request body. And then you don't have to worry too much about these other things for now. At the moment it's just going to be the request body, the API key, and the endpoint. So what we want to do, come back over to Make, and this is the automation we set up earlier. We're just going to get rid of this response webhook for now. And we've got this webhook here. This is the webhook that's going to connect to the Slack API. So we want to keep that. And what we're going to do is add another module, an HTTP module. We're going to go there. We're going to go to make a request. And then this is what's going to make the request to the Relevance AI agent to trigger it. So the custom webhook will be triggered when the, when a message is sent in the Slack channel. And then it's going to trigger this HTTP node, which is going to trigger the Relevance AI agent. So what we do in the URL section, we come over to here, this endpoint, we want that. So we copy that, paste it into the URL. The method, we're going to change that from GET to POST, and then we want two headers, the first one being for the name Content-Type, and then the value is going to be application/json. And then the second one is going to be this authorization, and then for the value, this is where we're going to put our API key. So if we come back over to Relevance, I'm quickly going to generate an API key and put it in there. And then for the body type, we're just going to click on raw. Content type is going to be application/json. And then the request content, I'm going to come back over to Relevance AI. You can see this request body, we're going to copy that. Paste it into there. And then the content, what we're going to do is we're going to get rid of this hello for now because we want that content to be the message from the Slack channel that this custom webhook is going to pick up. So what we need to do is we need to activate the webhook by just sending a message in the Slack channel, be able to get that content variable so we can put it in there. I'm quickly going to turn this off so we can do it manually. Come back over to our Slack channel and just say Hey. And then when we run this, you can see, yeah, there's data. So it's obviously picked it up. We're going to get the error, which is fine. But when we come over to the request content, you can see now we have all of these different variables. And we want the variable to be text because you can see that's the hey that we inputted. So that means whenever a user inputs anything, it's going to be saved under the text variable, which is what is going to be sent to the Relevance AI agent. So we hit text. And so what we're going to do now is we're going to send an actual request through the Slack channel. We're going to try and get the agent to actually write something and see if it goes through this Make.com scenario and actually triggers the agent. So if we come back over to the Slack channel, let's say, write me a piece of content on— so I've just said write me a piece of content on AI. We're going to send that. Come back over, and we're doing it manually for now just to check that it's all working. But obviously when we go to actually use this, these will all be turned on so that it's going to instantly receive the data. But for now, we just hit run once, use existing data, and obviously a slight little error there. Just accidentally added a colon to the end of the content type. So it just goes to show that you've got to be really particular with these things. Little things like that can trip it up. So we're going to try this one more time, use existing data, and then yeah, you can see it's sent off the request to the Relevance AI agent. And then you can see it has triggered the Relevance AI agent. And now the tool isn't working because we haven't set up that AI block yet, but the Relevance AI agent has been triggered, which is what we want. So now I'm going to show you how to actually get the content that the Relevance AI agent generates and put it back into the Slack channel so it replies to the message that you send inside the Slack channel. So we're going to go back over to that content generator tool, and obviously the reason why the tool didn't run was because the API step wasn't set up. So that means the whole tool just doesn't run, it doesn't actually generate an output. So what we're going to do is go over to Content Generator tool, hit Edit, and now we need to set up this API block. So what we're going to do is come over to Make, we're going to create a new scenario, and for this one it's again going to be a webhook. So same deal as before, custom webhook. Let's give it a name. So we've said content generator response, hit save. Again, we're going to get this URL, which we copy, come back over to Relevance AI. We change the method from GET to POST, and then the URL, that's going to be the webhook URL that we just got. And then the body, we're going to select raw. And then the content that's going to be in this that we're going to send to the webhook is going to be the variable for the step before, which I haven't actually named. So I've got to go ahead and name that now. I've just called it content generated. And then as you can see, yeah, so in the actual body of what it's going to send to the webhook is going to be the content generated by that LLM step before, which is good. And now what we need to do now, similar to before, is we need to like— is we need to sort of verify this webhook. What we're going to do is just going to run this whole tool. I need to put in the user input, so we can just say— so just say copy on AI, hit run tool. This should create a 150-word piece of copy on AI. And then, ah, as you can see, we didn't need to verify it. So that's just worked instantly, it's accepted it, which is perfect. So yeah, now it should say successfully determined. So that's good. So now what we do is we add another module, and this one is going to be a Slack module. So we type in Slack, create a message. Now you will need to connect this to your Slack account if you've never used this before. It's really, really simple though, you just click on it, it'll ask you to confirm or deny, take you through a little process. It's really, really simple. And then enter a channel ID or channel name. We're just going to select this from a list. Public channel, yep. The public channel name is the Content Generator Agent channel that we made before. And for the text, it's going to be the value from the webhook. So we can hit OK. Then we're going to come over to Relevance AI, and we're just going to test this one more time with the exact same prompt. And see if it sends it to the Slack channel. Okay, so this has all worked. It said accepted. So we're going to come over here. This is turned off so we can do it manually again just to make sure. So it's received the data, use existing data, and then as you can see, it looks like it's sent it. So we come back over to Slack. Yeah, so it's now sent that message into the Slack channel, which is perfect. It's exactly what we want. So now we're going to try and do it all together. We're going to turn all of the Make.com automations on, and we're going to see if we can just send a message into this Slack channel and see if it replies. So I've just said to it, create a piece of copy on the benefits of AI. We're going to hit send. And now I have found that this takes a little minute, and there's no type of loading animation, so you don't actually know if it's working. So what we're going to do is we're just going to wait a sec, and we're We're going to see if it loads. It should load perfectly, but we're just going to make sure. Okay, so it took a little minute. I realized it was a stupid little thing. I hadn't hit save on the Relevance AI tool. It's obviously just simple things like that can really affect the flow of it. But as you can see, it's now generated it, and it generated it pretty much instantly after I tried it again. So we're just going to do the exact same prompt again and see how quickly it generates it this time. So yeah, as you can see, it's pretty much generated it. Within the same minute, 10:31 to 10:31. So yeah, that's worked really, really well and super effective and just really, really easy way of doing it. So that's a really, really simple way of integrating Relevance AI into your Slack channel. Could be really helpful for, as I said at the start of the video, sort of web scraping or getting information about your business or upcoming projects or things like that, all within Slack. So if you have found this video helpful, I'd really, really appreciate if you subscribed and follow along with the journey. And if you are interested in integrating AI into your business, then I've put a link to website below, and you can book in a call, or you can just email me directly with any questions that you have on potentially integrating some of these systems into your business. And if any of you watching have any ideas for videos that you'd like to see, then just drop them in the comments, and I'll try my best to do them.
More from Rory

Dynamic Workflows are crazy. This is how to actually use them (beginner's guide)










