For most of my career, I treated product specs like sacred texts. They were proof I’d done my homework. That I’d thought through every scenario, captured every assumption, and documented every possible outcome.
But somewhere along the line I realised that product specs don’t always bring people together. Sometimes they divide them.
They’re long, formal, and let’s be honest, often ignored. Everyone reads them through their own lens, nods along in the meeting, and then walks away imagining something slightly different. Two weeks later, reality bites: “Oh, that’s not what I thought we were building at all.”
Writing a spec feels like the responsible thing to do. But that doesn’t mean it’s always the effective thing to do.
From Documentation to Demonstration
Lately, I’ve started doing something different.
Instead of writing down what I think we should build, I’ve started building what I think we should build; quickly, roughly, and with help from AI tools like Lovable, Bolt, and Magic Patterns.
What used to take a team days now takes a few hours on a laptop. I can sketch a flow, generate screens, connect them, and have something testable before the end of the afternoon.
And that changes everything.
Because a prototype isn’t just an artefact, it’s a conversation starter. It’s a way of thinking in motion.
When you show people something they can actually click and critique, they stop interpreting and start reacting. Suddenly, alignment isn’t theoretical, it’s tangible.
The Sampler Moment
It reminds me of what happened in music production during the 1980s.
The arrival of the sampler revolutionised how music was made. You no longer needed expensive studio time or a music degree. You just needed curiosity, creativity, and taste. People could loop, remix, and experiment their way to hits that sounded like nothing before them.
AI prototyping tools are doing something similar for product people.
They lower the barrier to exploration and make the creative process fun again. You don’t have to wait for engineering capacity or the next sprint cycle to explore an idea. You can tinker, test, and iterate on a Tuesday night while Netflix hums in the background.
The Taste Gap
Of course, democratising creation doesn’t automatically lead to quality.
AI can generate a thousand variations of an idea, but only human judgement can tell you which one’s worth pursuing.
Execution is getting cheaper by the day. But judgement, empathy, and taste remain the expensive parts of product management.
That’s what makes this shift so interesting: AI isn’t replacing what we do, it’s amplifying the parts of our job that matter most.
The New Spec
So maybe the best way to write a spec is not to write one at all.
Maybe the prototype is the spec. The shared artefact that gets everyone aligned faster, cuts through the noise, and surfaces what really matters.
Because when you start from something tangible, you invite challenge, clarity, and collaboration. You move faster, learn earlier, and waste less time.
Marty Cagan summed it up perfectly:
“If we can prototype and test ideas in hours and days… it changes the dynamics, and most important, the results.”
He’s right. Once you’ve worked this way, it’s hard to go back.
At hedgehog lab, we’re using AI to rethink how we prototype, not as a final deliverable, but as a thinking process. It’s helping teams make better decisions, faster. And in a world moving as quickly as ours, that feels like the real revolution.
The future of product thinking is already here.
At hedgehog lab, we’re exploring how AI can make creativity more collaborative, inclusive, and fast-moving.
