.png)
When I was a kid, there was a boy in my class, whose father was a real tinkerer. I'll call him Tom. He was working as a chimney sweeper and during the day he was as far from technology as possible.
But during weekends, Tom was a real builder. He collected used PC parts and built his own PC and all sorts of trinkets. He was really enthusiastic about technology but he wasn't making much while taking care of family of four.
Tom was building, learning by doing. Tom wasn't a kind guy and he was difficult to be around, but he was one of the most resourceful people I've ever met.
He was, what I call a quiet innovator. Not for fame or glory but to solve problems for his own family, with the resources he had at his disposal.
What I really love about building Promptmaster is that we're full of these quiet innovators. All working to ease the immense workload we're all dealing with and the potential of technology to lighten this load—a prospect made real by our students like Mike.
How Mike saved 84% of content creation time
Mike works as a nutritionist and in his business they're creating a lot of content for their clients. They call this a care journey. Mike reached out to us in April to help him solve a few problems around this but even I couldn't dream of where he would get in just 4 weeks.
It all started with a few simple issues. He was using Make to build super simple AI agents that would get data from their own systems (stored in Google Sheets), pass through OpenAI to generate articles and then save them to Google Docs. He was playing around several steps of refining content to make sure it's good quality (applying Conversation Design principles)
Instead of setting out to build a Death Star from the get go, he started small, which was smart. Within a week, he was seamlessly generating different types of content, articles, text messages, but the process became really complicated really soon.
During our second call I showed him a few tricks on how to structure OpenAI modules within Make to generate content dynamically while simultaneously improving output quality by a lot.
Now it wasn't just about creating articles. Oh no, he decided he'd want to automate EVERYTHING about content creation. So we discuss how he could do that for himself, building an entire system of AI agents powered by Make and OpenAI that generate articles directly into their own content management system.
Then today we had another session where he showed me what he built. When I saw it I knew I had to show it because it made me feel so proud. Within four weeks, he went from having trouble setting up a simple OpenAI module in Make to AI agents entirely personalizing the care journeys for their clients.
But it was a bit limited as it could only cover a certain type of care journey, with 6 different blocks of communication. So today we built the basics of how he can scale it to any kind of care journey.
The best part about this is that the assistants and nurses who were spending 2 hours on writing an article about blood pressure now generate several plus a bunch of hyper personalized text messages within less than 20 minutes. That's an 84% decrease in worktime without sacrificing quality.
You don't need to build an app
So I was naturally impressed, but when he showed me how his entire team operates the agent I was flabbergasted. Did he use an app? ChatGPT? Built his own thing? No.
He turned Google Sheets into a Control Center:
It's set up in a way that when a nurse changes status (the colored drop downs), the AI agent gets triggered. Basically different statuses trigger different webhooks that in turn send a different set of prompts to OpenAI.
This is great because their team doesn't have to use a new software. The entire cost of running an AI agent like this is nominal. You can create thousands of pieces of content for less than the cost of a ChatGPT Plus subscription.
Later on we went into optimization challenges of how to use different Make modules to make sure it's scalable to his needs.
Before Promptmaster Mike was like everyone else: spending countless hours on finding the right AI tools on aggregator websites, but none of them were cutting it.
Then he realized he could just DIY the whole thing if he knew how to do it. It's faster, easier, more reliable.
Mike is a quiet innovator, just like Tom was. He is solving problems for his own people, his business, his family. You don't need to spend thousands on AI software or pay even more to developers.
Oh and one more thing: Mike has ZERO engineering background. As I said: he is a nutritionist.
How to build your own AI agent
You can be like Mike. You can build your own AI agent within 4 weeks from start to finish. Even if you're starting from scratch.
No tech skills needed.
No coding required.
We have 12 year old kids, busy dads and 80 year old grandmothers in our community. If they can do it, you can do it too.
If Mike's journey is something you'd like to experience, consider getting our Prompt Master AI Course. It's not just a course, starting with a free 1 hour consulting session to create your plan on what to build.
Then you'll join a community of 1,200 highly trained experts and I'll be there for you all the time with unlimited DM support.
Talk soon,
David
PS: Some of our students are sharing the agents they build with Make in our community. Currently we have 18 different agents uploaded to our Resources library. You'll get that all for free once you're a member.
