AI, prompted correctly, can help generate marketing plans that hold up in practice. The method: simulate a 100-person focus group split into 20 breakout rooms of 5, gather their deliberations, then merge everything into unified recommendations – distilled into a one-pager and a longer 5-20 page report.

See a sample – WithCoverage (unaffiliated demo). Message me for the full report.

Model 100 fictitious people in AI, each with a variety of skills. Break them into 20 random groups of 5. Have them independently discuss a random unaffiliated company (Reevo) twenty times. Collect those comments and create a marketing plan out of it.

Many AI-generated marketing plans are throwaway junk – the fix isn’t a better tool, it’s a better prompt. Rather than asking an AI to just hand over recommendations, I get it to role-play an entire focus group: 100 simulated participants weighing in on the company. But they don’t all talk at once. I break them into 20 small clusters of 5, shuffled at random, and let each cluster debate independently before pooling everyone’s input into one synthesized set of concrete, actionable recommendations.

What comes out the other end is a report running 5 to 20 pages, but I never stop there. I always boil it back down into a tight one-pager afterward.

Case in point: the 1-page summary of Immigify’s marketing plan – a longer writeup exists too, just ask. Absolutely no affiliation with the company, just one picked at random to try out my new mousetrap.

Here’s something I didn’t expect: AI actually gets good at helping with marketing strategy once you stop asking it questions and start giving it a scenario. My trick is simulating a focus group – 100 fictional participants, split randomly into 20 breakout rooms of five – all debating the company from different angles. Each room deliberates on its own first. Then I pull everyone back together and have their combined input distilled into a unified set of recommendations. What comes out the other end is a report somewhere between 5 and 20 pages, plus a condensed 1-pager for anyone who wants the short version.

Let me show you how it plays out, using a company I have zero connection to: Adaptive Insurance. Bigger report available upon request.

Turns out AI can help produce marketing plans that don’t feel like generic filler – you just need to prompt it the right way. My breakthrough came from ditching direct questions entirely and instead simulating a 100-person focus group hashing out ideas about the company. I break that crowd into 20 breakout rooms, five people per room, randomly assigned, and let each group hash things out independently. Then everyone reconvenes, and their scattered input gets distilled into one clear set of recommendations. The output: a report anywhere from 5 to 20 pages, plus a tight 1-page summary.

Case in point – here’s the process run on a company I have absolutely no ties to: Norm AI. The longer report is available by request.

AI can generate genuinely useful marketing plans, but only with the right prompting. Through trial and error, I found that framing the task as a simulated 100-person focus group discussing the company works far better than just asking for answers directly. I split the group into 20 random breakout rooms of 5 people each, let them deliberate, then reconvene everyone to synthesize their input into a set of recommendations. From there, I generate a 5-20 page report, followed by a 1-page summary.

To show how this works, let me demonstrate on a random organization I have zero affiliation with: Handspring Health. Longer report available by request.

AI turns out to be surprisingly capable at generating marketing plans that are actually useful — but only with the right kind of prompting behind it. Through a fair amount of trial and error, I discovered that framing the task as a simulated 100-person focus group, all discussing the company in question, produces noticeably better results than simply asking the AI for answers outright. My process works like this: the 100 participants get split into 20 randomly assigned breakout rooms of 5 people each, where they deliberate independently, and then the entire group reconvenes so their collective input can be synthesized into one cohesive set of recommendations. From that synthesis, I generate a report ranging anywhere from 5 to 20 pages, along with a condensed 1-page executive summary.

To show how this actually works in practice, let me walk through it using a completely random organization I have zero affiliation with: Kyber. The longer report is available by request.