We’re a security startup where everyone wears five hats. This month, our product designer has become one of our most effective frontend contributors.

Last week, she fixed a responsiveness bug that had been bothering us since Christmas. It took her 10 minutes, and cost 76 cents.

She has no engineering background—no JS, no React. Just Claude, wired into her dev environment (iTerm2 + Docker), acting as a live-in copilot.

One prompt, one fix. It worked. And it shipped.

That’s not a fluke. She’s now added buttons, fixed spacing logic, tuned margins—stuff that would've gone into our backlog as Jira tickets. Things that typically get deprioritized because they’re small, but still matter.

Now they get fixed on the fly.

This is a major deal for a startup of our size. Fix a dozen pesky UI issues and you’ve seriously improved your user experience.

For transparency, here’s the spend:

Our total Claude spend for Allie as of last week.

Call it two weeks of occasional prompting, and a dozen changes shipped to prod. 

Total cost so far? $134.

She’s still not a developer. She still pings our engineers when something breaks, and everything she pushes gets PR reviewed like everyone else. And yeah, sometimes she borks the layout and has to start over. 

But the difference is that now, she tries. And most of the time, it works. She can now make a change in less time than it takes to create and delegate a ticket.

There’s a lot of noise about AI taking jobs. But in our case, it lowered the barrier to entry to make our product better.

I’m not saying this replaces engineers. We need more of them. Desperately. But AI changed who can contribute to our product directly. It lowered the barrier, and gave one more person the ability to make things better, faster.

In a week where every hour counts, that matters a lot.

‍

Overwhelmed by unstructured data? Datable helps you peel out noise from signal, and route the right data to its right home. Check out the sandbox here.