How I'm Training Claude to Build My Business

How I'm Using Claude Projects to Automate My Future

Today's podcast video is going to explain how to use Claude Projects to become your Make.com automation assistant. 

You may know that I've been doing a bunch of research using Notebook LM the past while. So I started studying all sorts of tutorials online so that I could get an idea of what everybody was doing with their AI automation right now.

So I studied what was going on with Make, what was going on with Zapier, what's going on with N8N, and I researched what are the cutting-edge tutorials for dealing with AI in tourism, AI for real estate brokers, for dental practices, and different business plans. AI for LinkedIn outreach. And the reason I'm using all of this is so that I can retire and not have to work anymore.

So I want to use AI to create the content that's going to make me the money. So then I knew a whole bunch more information. I was really feeling a lot more confident about where I am and what I'm doing, and I wanted to expand my studies more.

🔍 Inside the Early Adopters Community That Changed Everything

So I started studying all of the different AI agent schools to see what are they offering that might be interesting to me. And that's what led to today's discovery: how to use Claude Projects to become your Make.com blueprint creation assistant. So today I went and joined the community Early AI Adopters with Mark Kashef, and it's a fantastic place.

I already found some stuff in here that I'm going to show you right now that's just amazing. So there's going and AMA is happening in a few hours. I may check that out but it's a weekend so I'm going to be busy.

But one thing that I did go through, and you see there's updates going on that are live, I found this tips and tricks for "Know Thyself with AI," and I was just reading through it. And it says, put in this prompt: "Assume that I'm lying to myself. What am I refusing to acknowledge based on all our conversations together?" So I brought that over to ChatGPT and I asked it that.

💬 Claude, Tell Me the Truth: What Am I Missing?

And it says, "Doug, here's the straight talk. You already have what people are searching for, but you keep creating like you don't have it." And here I am doing all of that research and really building.

But the reason is I see the vision of how all of this stuff can work together.

And after watching Mark's video, I know he's got some of the answers that I need.

So I went in and I did this podcast that explains the entire thing of what I'm going to use Mark's knowledge for so that I can use Claude Projects to learn all of the best Make.com blueprints.

📚 How I Plan to Train My Claude Agent with the Best of Make.com

So the second step in my process after I get that machine built is to go back into all of these super great AI schools that have the really good blueprints for Make.com. So I'm grabbing all of the best coding practices, and then I'll train my Claude agent on those best practices.

As an end result, I'll be able to ask the AI to provide me with a set of blueprints that is specific for a business—for a tourism business. So we can get the business details and their description. We can get their ideal customer.

We can put that in there and the AI will create a custom set of blueprints that are designed specifically for that business in that industry, specifically for a real estate business agent in that industry, specific for a dental practice or for whatever your focus is.

🛠️ Building a Community-Powered Claude Blueprint Generator

If you wanted to run your own AI agency, you could then join my community, use my machine to create those blueprint sets for your clientele, and then it's easy to automate whatever you're doing. So now, in order to get into this information, I'm going to start the automation.

I'm going to start the audio overview where I had Google LM create a podcast conversation that walks through and explains the entire process of using Claude Projects to become your Make.com blueprint assistant. And those are my steps that I'm doing next.

And I'm happy to have you along. If you're interested in joining AI Early Adopters, come and join me. This is only 195 people in here.

It's because they can't find the place. It's a fantastic place and I hope to bring in a bunch more people. I will get 40% commission for life if you go through my referral link down below and I would appreciate that.

🎧 Inside the Podcast: Claude and Custom Automations for Any Industry

I'll see you on the inside here. Now, here comes the podcast that explains why I'm in here, what I'm seeking to build, and that puts me on the front line. I've got to now go and do it. Thanks for watching.

Oh, and by the way, there is a PDF here that you can get all sorts of information that I studied up about Mark's public YouTube videos. Thanks again, and I'll see you on the other side of the podcast.

Welcome back to another AI Business Plans Deep Dive.

Great to be diving in again.

So, today we're really zeroing in on a specific plan. It comes from Mark Kashef. Yeah, Mark Kashef, he's got this really interesting idea about using Anthropic’s Claude, right? The AI model, Claude. Exactly.

🔧 Training Claude on the Best Blueprints in the Business

Using Claude to help out with Make.com automations. Okay. Make.com, that's the platform for building workflows, right? Using those visual templates. They call them blueprints—pre-built automation templates. Yeah. It's pretty powerful stuff already. But Kashif's idea adds another layer.

He's talking about using Claude to actually generate new custom blueprints for Make.com. It's a cool concept. The core idea is training Claude first. Training it on what exactly? On what Kashef calls the best Make.com blueprints already out there.

Okay, the best—that sounds important. What does that likely mean? Like the most efficient ones, probably. Yeah. High-performing, well-structured automations, ones that really work well. So Claude learns from good examples, you know, right? It absorbs the patterns, the logic of what makes a successful automation tick.

⚙️ From Learning to Generating: How Claude Builds Custom Solutions

That's step one, learning from the best. And step two is the generation part. Yeah. That's where it gets really interesting. Claude then generates a whole set of new automation blueprints. Not just copies though.

No, definitely not. He says they're generated to specification. Ah, okay. So custom-made. Precisely. And crucially for a business within a specific sector. That's the key bit. So it's not generic.

If I run, say, a small online bookstore, Claude could potentially create blueprints specifically for your inventory or maybe your customer follow-up process. Tailored. That makes sense.

🧩 Claude’s Strengths and the Bigger AI Automation Puzzle

And Kashef has done other work in this area, right, with AI and automation. Yeah, he's definitely explored using different tools—things like Gemini for data analysis, I think—and even ChatGPT for building teams of AI agents.

So he's looking at AI from various angles. Exactly. That experience likely shapes how he's thinking about this Claude project, how to pick those best blueprints and apply AI practically.

What's the potential upside for businesses here if this works? Well, potentially huge time savings, I'd imagine. Instead of building complex automations manually or fiddling with templates that don't quite fit, you could get something custom much faster.

That seems to be the promise. Faster deployment, less manual setup, solutions that are actually designed for your specific needs could really boost efficiency.

⚠️ What Could Go Wrong? Key Challenges in Claude’s Automation Pipeline

And you mentioned Claude and Make.com—these aren't super inaccessible niche tools, are they? No, not really. I mean, they require some know-how, but they're generally accessible platforms, which suggests this approach could potentially be used by a pretty wide range of businesses.

Okay, but it can't be that simple. What are the hurdles, the challenges? Good point. Well, first off, the quality of those initial best blueprints you train Claude on. Right. If the examples are flawed, the output might be too.

Exactly. Garbage in, garbage out, as they say. The training data quality is absolutely key. What else? The specifications you give Claude—how clear are they?

Ah, yeah. If you just say, "Make my business better," Claude probably won't know where to start. Precisely. You need to be really specific about the problem, the workflow, the desired outcome.

Otherwise, the generated blueprints might just miss the mark. So, clarity is crucial. Definitely.

🧪 You Still Need Humans: Testing, Refining, and Smart Deployment

And then there's testing. You wouldn't just deploy an AI-generated blueprint without checking it. No way. You'd need to test it, refine it, make sure it actually works in the real world.

Mhm. So it's not about replacing humans entirely. It's more like AI gives you a really, really good starting point. Augmentation, not replacement.

You mentioned Kashef exploring AI agents too. Like custom GPTs that can trigger Make.com actions. Yeah, that's a related idea. He's looked into using AI agents to actively start automations.

It's slightly different from generating the blueprints themselves, but it shows he's thinking deeply about how AI interacts with these platforms.

So different AI models for different tasks in the automation pipeline, kind of. Yeah. Using each AI for its strengths. Claude seems suited potentially for understanding that structured logic of automation blueprints and generating new ones based on examples.

💡 The Practical Path: From Training to Real Business Results

Okay. So for you listening, the real aha moment here, I think, is seeing a practical step-by-step concept. Right. It's not just theory. It's: train this AI on these examples.

Give it these specs. Get this custom output using tools that are actually out there. It makes the idea of AI-driven custom automation feel much more concrete.

And for anyone who wants to, you know, really get up to speed quickly on AI applications—the learner persona we sometimes talk about—this is a solid example. A specific problem, a specific AI, a specific proposed solution.

🔁 Can This Model Be Applied to Your Industry Too?

Okay, so let's quickly recap this deep dive into Mark Kashef’s plan. We looked at training Claude the AI on top-quality Make.com automation blueprints. Yep. The best ones.

And the goal is for Claude to then generate new, totally custom blueprints based on specific business needs, right? Tailored for particular sectors or types of businesses.

We talked about the potential benefits—faster automation, more relevant solutions—but also the challenges like needing really good training data and super clear instructions for the AI. And the essential step of testing and refining anything the AI generates.

We encourage you to think about this model: training AI on successful examples to create new tailored solutions. Could that apply elsewhere in your work or industry?

What are the best practices in your field that an AI could learn from to generate something new and useful? It definitely leaves us with some things to think about.

🤖 Big Questions for the AI-Driven Future

Yeah. Like as AI gets better at this kind of generation, how do we even define what the best examples are for training? And how do we ensure the AI's creations are truly robust and reliable out in the messy real world?

Some big questions for the future. Okay. Really fascinating stuff. Definitely.

Well, if you found this deep dive useful and want more on AI, business, and innovation, please do subscribe so you catch our future discussions. And don't forget there's a study guide PDF for this deep dive in the show notes. It goes into a bit more detail.

Yep. Check out the show notes for that PDF download. Thanks for tuning in and diving deep with us today for a conversation.

➤ Learn more in our Skool: https://skool.com/aibusinessplans