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.

Unlocking a genuinely valuable marketing blueprint from AI comes down to creative prompting. Rather than asking the AI for flat, direct strategies, I’ve had success simulating a virtual 100-member consumer panel to dissect a brand. I divide this massive crowd into 20 separate, intimate workshops of 5 people, let them debate the company’s strengths and flaws, and then merge their collective insights into an actionable strategy. This collaborative simulation feeds into a comprehensive 5-to-20-page master playbook, which I then condense into a high-impact, one-page brief.

To show you how this crowd-sourcing prompt functions in practice, let’s test it on an entirely random target I have zero connection to: DataQuest.

Most AI-generated marketing plans are forgettable — generic, surface-level, the kind of thing you skim once and never open again. The fix isn’t a better prompt asking for “recommendations.” It’s changing who’s doing the talking.

Instead of asking the model to hand me answers directly, I have it role-play an entire focus group — 100 simulated participants reacting to the company in question. But I don’t let them talk all at once. I break them into 20 small clusters of 5, mixed up at random, and let each cluster debate the company on its own first. Only after that do I merge all 20 conversations together, pulling out the specific, actionable ideas that surface repeatedly or stand out.

What comes out the other end is a real working document — usually somewhere between 5 and 20 pages — followed by a second pass where I compress all of that into a single-page summary anyone can act on.

Here’s a sample 1-page run using Essent as the test subject (no relationship to the company — just a good example to work through). Longer version is also available.