Brainstorm with AI without getting generic ideas

Exit

Lesson 2 · 8 min

Brainstorm with AI without getting generic ideas

By the end

Twenty ideas for a real problem, and two worth trying that aren’t on everyone else’s list.

  1. 1Learn
  2. 2Try it
  3. 3Human check
  4. 4Quick check

Learn · 3 min

Why the first ideas are everyone’s ideas

Ask for “marketing ideas for my bakery” and you’ll likely get a giveaway, a loyalty card and a seasonal special. Not wrong, just the most common answers, because an AI model predicts likely text. A competitor who asks the same question gets much the same list.

To get past the obvious, give it what the average answer doesn’t have: your limits, your history and your taste.

Give it your limits and your examples

Constraints sound limiting, but they’re what make ideas specific. And examples beat adjectives: “fun but not silly” tells it little, while the post that actually got customers talking tells it a lot.

Rough notes are fine. “The giveaway got plenty of likes but no new customers” is exactly the kind of detail that changes what you get back. Give it:

  • Budget and time: “no ad spend, two hours of staff time a week”
  • Who it’s for: “office workers within a ten-minute walk”
  • What worked before, and why you think it did
  • What flopped, and why
  • What’s off-limits: “no discounts, nothing that needs a new hire”

Ask for more, then ask for different

Ask for 20 one-line ideas, not five polished ones. Quantity gets you past the obvious, and cutting a list is quick.

Borrowing from other industries works well, because it has read about so many of them: a car wash’s monthly membership might become your bread subscription. Then push it somewhere new:

  • “Skip anything a competitor would think of in five minutes.”
  • “Give me three that feel slightly risky.”
  • “How would a gym, a bookstore or a car wash solve this?”
  • “Combine two of these into something new.”

You still choose

The AI can’t see what your competitors did last month or what your regulars are tired of, and it doesn’t know an idea is unworkable for a team of six. It generates options. Picking is your job.

Take the two ideas you like best and ask it to argue against them: “What would make this fail for a business like mine?” A quick critique catches the obvious problems before you spend anything.

Then ask for a first step for each, and try the cheaper one first.

Try it on your business

5 min

Pick a real problem you need ideas for this month: a quiet weekday, a launch, more reviews, a way to win back customers who’ve drifted off.

  1. 1

    Name one real problem or goal. The narrower it is, the better the ideas.

  2. 2

    Fill in the brackets with real limits and real examples, even rough ones.

  3. 3

    Run the prompt, then follow up once on the idea you like most, for example: “Give me three variations of idea 7.”

  4. 4

    Choose two ideas you’d actually try, and write down the first step for each.

I run [YOUR BUSINESS] for [WHO YOUR CUSTOMERS ARE].
I need ideas for [THE PROBLEM OR GOAL].
Constraints: [BUDGET, TIME, CHANNELS AND ANYTHING OFF-LIMITS].
What’s worked for us before: [ONE OR TWO EXAMPLES, AND WHY THEY WORKED].
What didn’t work: [ONE EXAMPLE, AND WHY].
Give me 20 ideas, one line each. Skip the obvious ones any business like mine would try first. Make at least five of them unusual or slightly risky, and borrow at least three from a completely different industry.
Then pick the three that best fit my constraints and say why in one sentence each.
Open ChatGPTOpen ClaudeOpen Gemini

Human check: before you use it

Quick check · question 1 of 2No grades

Why does “give me marketing ideas for my café” tend to return the ideas everyone gets?

Ask the Distillery

Answers come from The Distillery’s own pages, with sources.

If the Distillery doesn’t know, it says so and points you to a person.