What it reads
The signals it works from. Typically connects to Document storage, Chat, Support desk and Intranet.
Answers the internal questions your team asks all day from your own documentation, and tells you which answers are missing or stale.
internal tickets
to the Scan's evidence — before any scope
to a system in production, not a pilot
modeled targets · not client results · timings are CRFT's standard engagement
The answer exists, in a document, that nobody can find. So people ask a person — the same person, the same five questions — and that person becomes a lookup service instead of doing their job. Meanwhile the documentation rots, because nothing tells anyone which page went wrong.
Where the Scan startsWith your docs, the questions people actually ask and where the answers go stale.
Four lines, the same four on every system we build. Where they fall is what makes this one different from the last one.
The signals it works from. Typically connects to Document storage, Chat, Support desk and Intranet.
The calls it makes on its own, unattended, every time it runs.
Into the tools the work already lives in.
Anything it is not confident about goes to the person who owns that area, and the answer they give becomes documentation. Refusing to answer is a designed behavior, not a failure mode — a confident wrong answer about policy is worse than no answer.
The same four legs as every CRFT engagement — written for this system rather than in general.
We categorize a month of real internal questions. The distribution decides the build — a long tail needs different treatment from twenty questions asked constantly.
Scope is set by source sprawl. Documentation in four systems with three conflicting versions is a content problem before it is an AI problem, and we will say so.
Confidence thresholds and source citation ship on day one. It is allowed to be less useful in order to be trustworthy.
The gap report is the compounding part: it turns the questions it fails into the documentation it needs, so the corpus improves because the system is running.
Internal tickets deflected, answer accuracy on a sampled review, and documentation gaps closed.
Five inputs, a published price, a report in 48 hours. Decide with numbers — then own what gets built.