Every investor says the same thing: services don't scale.

We're a software company. We have a self-serve platform. Teams create tests from plain English, run them in CI, get reports. Pure SaaS.

But we also have FDE pods - Forward-Deployed Engineers who handle QA testing end-to-end for larger customers. They plan tests, verify results, file bugs, gate releases.

Sounds like an agency, right?

Here's what I've learned: the service layer isn't a compromise. It's the product lab.

Every week, our FDEs see patterns. Where the AI fails. Where customers get stuck. What "done" actually looks like for a VP of Engineering who just wants to ship without worrying, catch regressions early.

That feedback doesn't come from analytics dashboards. It comes from being in the workflow.

We take those learnings and bake them into Studio. The service makes the software smarter. The software makes the service more leveraged.

There's a debate happening right now: are agencies cooked? Can't Claude just do it?

Maybe for some things.

But for high-stakes work - where quality matters and mistakes cost real money - you need controlled, responsible AI-powered services.

Humans in the loop. Judgment. Accountability.

YC just published an RFS on this: AI-Native Agencies. Their take - AI lets you sell outcomes with software margins. Not hours. Not headcount.

That's the bet we made early @ bug0. Still early, but feels good to see the thesis validated.

Originally posted on X