A film producer I work with has made some movies you would recognize — I will not name him, but you would know the work.
When he is developing a new film idea, he has a brainstorming process that surprised me when he described it.
He takes the idea to ChatGPT first. Types out the concept, has a conversation, sees where it goes. Then he copies everything — the original idea, the back-and-forth, what came out of it — and pastes it into Claude. Same question, new conversation. Then takes that and brings it to Gemini.
Three separate sessions. Three different models. All on the same idea.
When I asked why, he was pretty direct about it: every model has different ideas and insights that the other models could not see.
He had found through experimentation that ChatGPT would catch certain angles, Claude would push the concept in directions ChatGPT had missed entirely, and Gemini would surface things neither of the first two had touched. Deliberately running the same idea through all three gave him a more complete picture than any single model could.
Why This Works
The AI models available today are genuinely different — not just in capabilities, but in how they approach problems.
ChatGPT tends to be strong on general brainstorming and broad ideation. It has been trained on a huge range of content and has a natural fluency in exploring possibilities. Claude tends to have a more analytical bent — it often pushes back, asks clarifying questions, and looks for structural coherence. Gemini draws on Google data and tends to surface connections that the other models might not reach.
For a film producer brainstorming story ideas, those differences matter. A good story idea needs to hold up to different kinds of pressure — emotional resonance, logical coherence, market relevance. Running the idea through three different editorial minds is a fast way to stress-test it from multiple angles.
This is what I think of as multi-tool native thinking. The best AI users do not pick one model and stay loyal to it for everything. They learn what each tool is actually good at, and route their work accordingly.
The Manual-to-Automated Arc
When he described his process to me, I saw two things simultaneously.
The first was how smart his underlying logic was.
The second was how much friction there was in executing it. Copying and pasting between three separate browser windows, reformatting the context each time, keeping track of what each model had said — it was a lot of overhead for something that could be systematic.
So I built him a workflow. He puts the idea in once. The workflow runs it through all three models, collects the responses, and returns three distinct perspectives in a single output.
Same result. One step instead of ten.
That is usually how automation decisions should work: you identify something someone is already doing because it genuinely adds value, then you remove the friction around doing it. You do not automate things people are not doing for good reason. You automate things people are already doing manually for good reason.
How to Apply This Without Building a Workflow
You do not need a custom workflow to try this. You can do the manual version yourself, and it is worth doing at least once to see if it changes anything for you.
Pick a problem you are working through — a strategic decision, a creative concept, a piece of writing you want to get right.
Take it to ChatGPT first. Have a real conversation, not just a single prompt. Then copy the full context and bring it to Claude. See what is different. Then take both to Gemini.
You will notice that the models disagree in interesting ways, emphasize different things, and sometimes surface angles you had not considered. That is the point.
Once you have done it manually a few times and confirmed it is actually useful for your work, then it makes sense to build or find a workflow that reduces the friction.
The principle behind what the film producer was doing is simple: for complex problems, multiple perspectives beat one. That is true whether you are talking to people or to AI models. The best ideas usually get sharper when they are pressure-tested from different angles.
AI just makes getting multiple perspectives faster than it used to be — if you know to ask for them.
