Thirsty Horses

Is there an AI prompt to give us breakthrough growth ideas?

Maybe, but you won’t like them.

Imagine if AI gave you the perfect growth strategy tomorrow – you wouldn't follow it. Because you wouldn't believe it.

Ask any consultant "What do you get when you lead a horse to water?"

A thirsty horse.

Because the best ideas sound bad at first.

Why?

Most founders have a mental model of their market that's 80% right. But the breakthrough ideas hide in the 20% they got wrong – which is exactly why those ideas feel crazy at first.

Here’s an example…

Two years ago, we worked with a SaaS for a certain type of freelancer. Customer interviews revealed their main job-to-be-done was generating structured proposals to impress potential clients.

Great insight, but they didn't test it for 18 months.

When they finally tested it in ads, they got good cheap leads immediately.

Why did it take them 18 months to test it? It contradicted how they saw themselves: as an "end-to-end solution," not a mere proposal tool. So it went to the bottom of the list.

Often the bottleneck isn't generating good ideas – it’s recognising them.

If you've been working on a problem long enough, the right answer is probably already on your list. There's a reason you're not doing it. And ChatGPT can't fix that reason.

AI can recommend ideas. But it can't make a horse drink – even if it's thirsty.

Simple next step

Before you write more prompts, open your team’s mental aperture with a “bad ideas only” brainstorm focused on your biggest blocker.

The “bad ideas” framing shines a light on ideas that lie beyond our arbitrary mental constraints. Most of those ideas will be crap. But you don’t need most of the ideas – you only need one.

20 Breakthrough Ideas

I documented 20 of best growth experiments of the past 10 years. Most require almost no traffic, budget, or tech. The PDF includes step-by-step instructions, examples and screenshots.