How the best AI creators use AI

Yesterday I shared a few of the bigger-picture lessons I took from a recent AI bot-building workshop I attended.

The focus of that post was on how to think about AI differently: as a process, a system, a mirror for your own clarity.

But what really stuck with me wasn’t just how these guys think about AI. It was how they use it.

When you watch high-level builders work, patterns start to emerge. Not in the tools they use (those will keep changing), but in how they structure their workflows and get leverage from their ideas.

Here are a few of the patterns I kept seeing that anyone can apply to get better results.

AI is a co-creator, not a magic wand

They didn’t rely on AI to do all the work. They paired it with human oversight. The model was often tasked with generating, structuring, or remixing, but a person was always steering the process, checking quality, and refining the result. This was especially clear in workflows that included a "content checker agent" or modular task chains to revise and improve outputs.

They built modular workflows instead of one-off prompts

Every speaker emphasized systems thinking. They structured AI use into defined steps. One example was having agents for transcribing calls, then analyzing research, then drafting content, and so on. Each step fed into the next. Others used sketching tools to map ideas and break tasks into discrete pieces the AI could act on more precisely.

They constantly captured, refined, and reused their own knowledge

Many described the habit of feeding in meeting notes, saved chats, research PDFs, and call transcripts to build centralized, reusable AI inputs. Over time, this became their second brain. They could reuse what they had already done to generate better and faster results later.

They organized their AI work around projects, not chats

They weren’t just talking to bots in isolated threads. They created persistent AI workspaces structured by project, where prompts, docs, outputs, and decisions all lived together. This allowed for smarter, more consistent output because the AI had context. And it kept everything organized for the human too.

Speed and clarity were optimized at every handoff

One speaker highlighted three invisible handoffs in AI work: input, processing, and output. Most people get stuck at these transitions. The pros were fast at getting data in, had prompt libraries ready, and minimized tool-switching or scattered thinking. They optimized not just for what the AI was doing, but for how quickly they could move through the process.

They used AI for thinking, not just tasks

A recurring pattern was using AI to help refine ideas. They asked it to take on different expert perspectives, break things down by principle, or explore edge cases. One speaker called this “remixing knowledge dimensions,” pushing AI to think across domains, not just answer surface-level questions.

They treated their prompts and outputs like assets

Nothing was throwaway. They saved good prompts. They saved great outputs. They refined what worked and built libraries of reference material. These weren’t just shortcuts—they were systems they could return to, adapt, or even package.

These patterns aren’t about fancy tools or hacks. They’re about thinking intentionally. They’re about getting organized, building systems, and letting AI amplify your process and thinking, not replace it.

If you’ve been trying to figure out how to get better results from AI, start there.

For me, it highlighted the need to create more systems around how I use AI. I’ve been using ChatGPT projects for a couple months, but I’m still a bit scattered in how I save, store and use my prompts and my data.

This is all new stuff, though, and the cool thing is that no one has it figured out yet.

So with a little experimenting and thoughtful application, anybody can become an expert.

If any of these ideas sparked something for you, if they gave you a new way to think about how you could use AI differently in your own work, I’d love to hear about it.

Hit reply and let me know what landed.

Nathan

When you’re ready, here’s how I can help

I build personalized AI Voice Bots that turn your own words and stories into ready‑to‑send drafts in seconds so you can hit Send on time and on brand.

Book a free 20‑minute Fit Call

I take only a few clients at a time. Even if we’re not a match, you’ll leave the call with some quick wins you can use immediately.