AI made unfinished thinking look finished. Superuser is building the pre-send layer that catches the gaps — in the numbers, claims, assumptions and caveats — before work reaches a client, board, boss or market.
AI adoption outran AI assurance
Every knowledge worker now ships AI-assisted output. Almost no one verifies it before it reaches a client, board or regulator. The drafting got automated; the checking didn't.
The moat is the referent
We don't grade opinions. Every check is tied to a source the user supplied — numbers, clauses, citations. Code proves the facts; AI only reviews the words. That's defensible and hard to fake.
Land in one wedge, expand across work
We start where the cost of a wrong number is highest — accountancy commentary — then extend the same engine to legal, research, consulting and sales.
We're early and deliberately so. The product is built, the verification engine works, and we're opening founding pilots with UK accountancy firms to lock in week-0 baselines and provable before/after deltas. We'd rather show you a real pilot than a vanity metric.
That discipline is the point: a product built on "provable, not made-up" can't lead with invented traction. The proof is the pilots.
The same engine that checks a management-accounts commentary checks a contract clause, a research citation, or a sales commitment. Each vertical shares the core — a verifiable referent and a conservative, human-in-the-loop verdict — so expansion compounds rather than restarts.
Talk to the founders
If you back category-defining infrastructure early, we'd like to show you the engine and the pilot pipeline.
investors@superuser.appThis page is information about Superuser, not an offer or solicitation to buy or sell any security. Any figures shown are illustrative.