One Piece of Work, Three Outputs

I need to produce 20 video guides for ka-do.

The obvious approach: screen record yourself doing each flow, talk over it, edit, upload. Repeat 20 times.

I'm not doing that.

As a solo founder, screen recording is the wrong job to be doing. Every hour I spend watching myself click through a UI and re-recording because I stumbled over a word is an hour I'm not spending improving the product.

So I built a pipeline instead.

Each video is a Playwright test. I write a script with beats: the moments in the flow that need to land for a new user to get it. Chatterbox voices it. The whole thing runs automatically and produces a finished video.

What this gave me that I didn't expect: I'm now an observer, not a performer.

When I'm recording myself, I'm focused on not making mistakes. When the pipeline is running, I'm watching the product. Does that transition feel obvious? Does that label make sense? Would someone who's never seen this before know what to do next?

The pressure of making a guide that actually works, one that has to hold up on YouTube and as social content, surfaces UI friction that normal QA never catches. You only notice that a button label is ambiguous when you're trying to narrate why someone should click it.

And the same Playwright test that produces the video is also a test. The video guide library and the automated test suite are the same thing.

One piece of work. Three outputs: content, QA, UX review.

That's the right use of solo founder time.

This same review harness is what eventually answered a much bigger question in When Is a Product Ready to Launch? The same discipline applies to writing, not just video, which is what Eight Posts That Were Secretly One Post is about.

Steve Drew is building ka-do, a calm task and planning tool for busy professionals.