Here's how the best idea in a pitch almost didn't happen.

We were working on Visit Pittsburgh. The brief was what you'd expect. Showcase the city. Make people want to go. The usual tourism playbook of pretty shots and a tagline that sounds like every other tagline.

And then somebody in the room — I honestly don't remember who — said the dumbest thing I've heard in a brainstorm in years.

"What if Mr. Rogers and Wiz Khalifa made a song together about Pittsburgh?"

The room laughed. Obviously. A dead children's TV host and a rapper. About a city. As a tourism campaign.

No sane person greenlights that.

The dumb ideas are the good ones.

I've been doing this long enough to know that the ideas that make a room uncomfortable are usually the ideas worth chasing. The safe ones — the mood boards full of golden hour skyline shots and "discover your next adventure" — those die in market. Nobody remembers them. They exist to make a client feel like they did something, and then they vanish.

But a dead kids' show host duetting with Wiz Khalifa? That's sticky. That's the kind of thing people text to their friends. That's the kind of thing that lives on the internet way past its media buy.

So instead of killing it, we leaned in.

Prototyping the impossible.

Here's where it gets interesting. Ten years ago, this idea dies in the brainstorm. You can't pitch a client on "imagine this thing that doesn't exist yet." You'd need to produce a demo, and producing a demo of Mr. Rogers singing with Wiz Khalifa would cost a fortune and take weeks. The pitch would be over before you started.

But we had AI tools. And that changed everything.

ChatGPT wrote the lyrics. I'm not going to pretend the first draft was Shakespeare, but it captured the tone — warm, nostalgic, a little bit funny, unmistakably Pittsburgh. We iterated on it three or four times until it felt right. The whole thing took maybe an hour.

Midjourney generated the visual concepts. Mr. Rogers in his cardigan, Wiz in his element, the city behind them. Not photorealistic — stylized, almost like an album cover from a universe where this collaboration actually happened. It looked like something you'd want to share.

Kling handled the motion. Took those stills and gave them life. Subtle animation, lip sync, the kind of thing that three years ago would've required a full production team and a six-figure budget.

In 48 hours, we had a prototype of a campaign that would've taken months the traditional way. Not a deck. Not a mood board. A living, breathing thing that a client could watch and feel.

The point isn't the tools.

I know what you're thinking. Cool, another AI case study. Another creative director bragging about prompt engineering.

That's not the point.

The point is that the idea came first. A weird, uncomfortable, shouldn't-exist idea that nobody in the room was sure about. The AI didn't generate the concept. A human did. The AI just made it possible to show it instead of describe it.

That's a massive difference.

The best creatives I know aren't using AI to come up with ideas. They're using it to keep ideas alive long enough to prove they work. Because in this business, the gap between "that's a weird idea" and "that's a dead idea" is about six seconds of silence in a conference room.

The best ideas don't come from tools. They come from people willing to say the dumb thing out loud.

AI closed that gap for us. We went from "huh, that's weird" to "holy shit, look at this" in two days. And suddenly the weird idea wasn't weird anymore. It was undeniable.

What I actually learned.

Every pitch has a moment where someone suggests something that makes the room tense up. The career move is usually to let it pass. Nod politely. Move on to the safe option.

Don't do that.

The tension is the signal. If everyone in the room is a little nervous about an idea, it probably means it's actually different. And different is the entire job. We're not hired to make the same thing everyone else is making. We're hired to make things people notice.

Mr. Rogers and Wiz Khalifa should not exist in the same sentence, let alone the same campaign. That's exactly why it works.

The ideas that shouldn't exist are the ones worth fighting for.

One more thing.

If you're a creative and you're not using AI to prototype your wildest ideas, you're leaving your best work on the whiteboard. The technology isn't the creative. You are. But it's the best thing that's happened to keeping dumb ideas alive since the invention of the napkin sketch.

Try it. Chase the idea that makes you nervous. Build the thing that shouldn't exist.

Pretty simple really.