[00:00:00] Speaker 1: Running my own business by myself means I have to handle all the small, repetitive tasks that pop up before I can get to the work that really matters. Automation workflows can take a lot of that work off my shoulders, especially if they can connect across different applications. But I've always struggled with these because inevitably there's always some piece of information from one app that doesn't really connect to a specific field that the next app is expecting. What I really need is a step in the automation workflow that can figure out how to bridge those connection gaps, something a little more intelligent. I'm Nick, and today we're partnering with Make. With the new version of Make AI Agents, I can easily build visual automation workflows that have actual decision-making built in. AI Agents and the new GridView make the whole thing pretty transparent. You'll be able to see how each step is working and make adjustments as needed with no coding required. So here's my situation. I run training courses for clients as a freelancer. Client requests come in mostly through email, and I manage my projects using Trello. Whenever I get an email from a new client, I need to add them to that Trello board manually. I need a system that will pick out those project requests, ignore unrelated emails, then route them to my Trello board. Now, if you need something that just automates the flow of data from one app to another, Make can certainly help you with that. But I don't want to just pass data. I'm actually going to cook up two separate workflows to streamline this whole thing with AI Agents making decisions. If you've used Make recently, you may have seen a button for Agents in the left navigation, but don't worry if you're not seeing that anymore because Agents are now completely integrated into Scenarios. Scenarios are the workflows that you build in Make. So I'm going to create a scenario, and this is the visual canvas where you can build steps and workflows that include Intelligent Agents. And Scenarios always start with a trigger. I'm going to choose Gmail and tell it to watch my emails. I would bet that this is the most common scenario trigger, but we'll see a different trigger when we see that second workflow. And if this is your first time connecting to a third-party service, you need to set up that connection and approve some permissions. Now that I'm signed into Gmail, setting up the trigger is pretty straightforward. I'll leave this set to Simple Filter with the folder set to My Inbox. It's basically watching for new emails in my inbox. When you save an email trigger, you have to decide if the scenario will run against all of the messages, messages within a specific date range, or just new messages from now on. That's the option I want, so I'll choose that. Now to really understand how these scenarios work, you need to know what information is being passed from one step to another. So it's a good idea to run the scenario even with just a trigger module in place. Since this trigger is looking for a new email, I'm gonna go ahead and send a test email to myself to give it something to read. Now that there's a new message in my inbox, I'll run this scenario. And when that's done, you can click the button above the trigger module to see the information that it got from that step. These are the information fields that it found from my email, and I can pass any of these fields from one step to the next in the workflow. To add our next step, I can click the plus button on this module to add a new module. I could pass information from that trigger to something like a Google Sheet, an Airtable database, directly to Trello, or there are thousands of other apps that you can connect here. But for my next step, I'm going to use an AI agent. You can do a quick search for make AI, then choose make AI agents. I'll choose the option to run an agent, then do some configuration. This is where the decision-making will happen. We'll need to create the connection to make's AI agent system. You can use different AI providers, but I'll stick with make's built-in provider. Then you can choose an AI model. This won't be too complicated, so I'll just choose GPT-5 medium to be safe. Then the instructions field is where you tell the agent exactly what you want it to do. This is a lot like the prompt you would write in an AI chat assistant. When I click that field, I have the option to add any of the fields that were passed from the original email message. Every email will have a different sender name, sender address, and everything else. You can use these fields to stand in for that information in your prompt. So I want this agent to screen my incoming email messages and identify only new messages that are not part of an existing thread. I don't want it to keep passing every reply from an ongoing conversation into my Trello board. Next, I'll tell it if there is a new message requesting information about training services, add a new card in Trello. So I'm asking the agent to read the email and understand whether it contains a training request. I'm asking the agent to make that decision. I'm also referring to a specific board and a specific list in my Trello account. Next, I'll add in the card name include the sender's name. Okay, so the sender's name is one of those fields that's being passed from the original email. So I'll drag in the name field from that email. Now you don't have to drag these fields into the instructions prompt. I just wanna be as clear as possible. However, you do have to add any fields that you want the agent to use to this input field. This defines the incoming data from the previous step for your agent to process. If you don't include important information here, then your agent will not be able to analyze it. Ultimately, I will be referring to the thread ID, the email address and the full body text in my instructions. So I'll just add those now. It will definitely need to refer to the full body text since I'm asking the agent to read and understand the incoming email message. And back in the instructions, I'll tell it, if you can identify the company name from the email message, add that to the card name, then add the thread ID. So the new card that will be added to Trello will have a name field at the top. And I'm asking the agent to put the email sender's name and the company name into that field. And I'm also asking it to include the thread ID. That identifies the specific email thread that this information came from. I don't need that right now, but it will set something up for when we get to my second workflow. And I'll also tell it to add some content to the Trello cards description. Now, before you save and close this, you should stop and think about what the agent needs to complete these instructions. According to the instructions I gave the agent, the next part of my workflow requires the agent to do something in another app, in Trello. So my next step is to give the agent a way to connect with and work in Trello. With that in mind, I can save this. And this time I'm not going to click the plus button on the right to add a new module. AI agents have a plus button on the bottom to add tools. I need to give the agent access to tools to work in other apps. In this case, I need to search for Trello, select that, and then choose a specific Trello action. The action I need is for it to create a new card in Trello. I'll need to sign into my Trello account and give it access. You can change the tool's name and description if you want. And in Trello, you have to tell it which list to connect to, but it has this nice option where it can let you choose the specific board and list from your account using a menu. So in my freelance projects board, it will add a new card to my list called new leads. I'll skip some of these optional fields. And this next step is cool. I need to set the name for the new Trello card. But remember, I asked my AI agent to put the name of the person who sent the original email into the card name. So I'll set this field to let AI agent decide. The agent has its instructions. It knows what to put here. And I did ask the agent to also add something to the card's description. So I'll set that to let AI agent decide. I don't need anything else here, so I'll save it. So this is my complete scenario. But you can link together many more modules and tools. But let's test this as it is. In my email, you'll see I've received two new messages. The message from Ava is a request for training. And the message from Liam has absolutely nothing to do with training. My agent should be able to sort through these. Let's look at Make and my Trello board side by side as I run this. The button at the bottom will run the scenario once. It checks my email. Then I see the agent is thinking. Then we're waiting for the Trello tool to finish. And we see a new card has been added to my Trello board. But just one, the agent successfully identified one message that relates to a training request, and it ignored the message that had nothing to do with training. That's exactly what I wanted. Of course, you should not expect your scenario to work perfectly the first time. There's a button above each step in the scenario that you can click to see a summary of that step. If something didn't work the way you expected, you should be able to find some information here, make some changes to your configuration, and retest until it works right. Now, since I have a scenario that works, I want it to run automatically. I'll save it. And there's a switch to set it to run every 15 minutes. Or you can click that and set a different interval. Now, let's take a quick look at another workflow. We're not gonna go through the whole process, but I wanna show you what it does. In my scenarios, I've already set up scenario two. The trigger is a specific action in Trello. When I move a card to a specific list in Trello, it will call the AI agent, which will write a response to the original email asking the client to schedule a meeting. Remember how I added the original email's thread ID to my Trello cards? This agent will find the thread ID in the card that moved. And I also gave the agent a description of the message I want it to write. And you can see I gave the agent a tool so that it can connect to my Gmail and reply to a message as instructed. Once again, let's look at make and my Trello board side by side. First, I'll move that new card to the contact list. This is the action that the scenario is looking for. Then I'll run the scenario. And it looks like everything worked. Now the original person who sent the request will get a response inviting them to set up a call. So I've set up two scenarios in make that work together to completely streamline new client communication. Now, if you just set up one scenario, then you've got everything you need. But if you set up multiple scenarios with complex flows that connect to different apps, that can be a lot to keep track of. If you do set that up, I think you'll love the grid. Choose the grid from the navigation bar. And the grid shows all of your connections across all of your scenarios. These three modules here represent a scenario that I was playing with earlier. But this little network over here represents my other two scenarios. They're connected because they both connect to my Trello workspace. I can see that my scenario two triggers the workspace and scenario one reads it. If there were any connections that weren't working properly or need your attention, you'd see those listed and you could troubleshoot them visually here. You can switch to different layers to see which steps are sending the most data over the internet or which are using the most AI credits. And if you are part of a team working together in make, this can help you visualize and troubleshoot connections across your team. So you can get started for free at make.com. You can start building powerful scenarios even with a free account. And it scales up for larger teams and enterprise users. As for me, no more missing client emails, no more sifting through my inbox to make sure I'm not forgetting something. And I don't even have to send that introductory email to schedule a first call anymore. The AI agents in make have helped me offload a lot of those administrative tasks that were taking up my time. To keep learning and for more videos like this, make sure to subscribe to this channel.
We’re Ready to Help
Call or Book a Meeting Now