UXSense

At EyeQuant we predict what people will see before anything ships. UXSense is about the other half of the question: what people actually do once it has.

It came out of a problem I've run into more than once. A feature technically works, but after a release returning users suddenly behave differently, and nobody notices for weeks. The dashboards look normal, nobody has time to watch hundreds of session recordings, and the hesitation, retries, and abandoned steps only become obvious much later.

UXSense compares how people behaved before and after each release, using recorded sessions, and turns the differences into a release report: what changed, where friction went up or down, and how much it's likely to matter. Every number comes from real sessions rather than being generated by AI.

It also works in the other direction. Because it knows which parts of a product people rely on most, it can check a pull request before it merges and flag changes likely to break a path people depend on, or move something they've built muscle memory around.

That's becoming more important as more software gets written by agents. UXSense can generate a BEHAVIOR.md file describing the parts of a product people rely on, so coding agents can read it before they start changing things.

I'm building it as founder, across product, design, and engineering, and it's still early.

Shipping faster is only useful if you can tell whether what you shipped actually works for the people using it.