Discovery interview

Purpose

Discovery says what exists. The interview says how bad it is, who owns it and by when — one estate domain at a time, with the tenant's own evidence on screen while the customer answers.

Steps

1. Open Console › Assessment workspace. The header reads n of m domains carry an analyst score.
2. Clear the unrouted pile first if there is one — see below.
3. Pick a domain card and press Assess this domain (Review this assessment once it is scored).
4. Read the evidence drawer aloud before asking anything. It is the answer to "where did you get that?".
5. Score the four rubric dimensions 1–5. Anchors and corroboration rules are on the assessment rubric.
6. Fill the dates and the owner: Deadline to meet, We finish by, Fix available from, Roll-up weight, Accountable owner, Mitigations in flight (one per line).
7. Why these scores is required once any cell is scored — a board figure with no recorded reason cannot be defended later. Board narrative is the sentence the board reads about this domain.
8. Press Save assessment. Saving writes your inputs; the deterministic engine recomputes the component score and the board headline from them.

What the screen shows

Domain cards — each carries the same four things in the same order: what it is, its state (Discovered, Assessed only, Benchmark anchors — figures the sector pack opened with, which no analyst has signed — or Not covered), the two readiness figures with their provenance, and who assessed it. A domain this organisation has no record for says so and points at Console › Onboarding instead of linking to a form that would write nowhere.

Evidence drawer — one row per routed asset: Host · Algorithm · Expiry · PQC status · TLS grade · Source · Routed by, with Move when your role can reroute. Its empty state names the next step: Run a discovery scan from the Console to populate it, or See what every domain would take to cover.

Unrouted assets (n) — assets counted in Inventory that roll into no component, so they reach no readiness figure and no board score. Pick the domain each belongs to and press Move — or route it to no domain if it belongs to none. Arriving from the coverage map's link opens this panel first.

The stamp — three lines, always both figures:

• Derived NN% · live · n assets — round(100 × PQC-ready or hybrid ÷ probed) over the probed assets routed here. With routed assets but nothing probed it reads Derived — · live · n assets, none probed.
• Analyst NN% · typed — an analyst's judgement of this domain, not a measurement. Unset it reads Analyst — · not assessed.
• Component score NN · live · computed by the engine, with its basis and coverage. With nothing to score from it reads Component score — · not scored yet and the reason — a zero there would read as low risk.

Above the rubric: QubiTrust assessment rubric v1 — these anchors are QubiTrust's own wording for this PoV, not a regulator's scale and not a standard. Where the estate agrees with a cell, that cell is badged live · your estate agrees.

Provenance

Common failures

• A domain no connector can populate. The card says no evidence path — analyst-led and names what would be collectable. Say so in the interview; the score is a judgement with no corroboration, and the stamp records that.
• available:false answers. "We have no PQC option from the vendor" is a typed assertion. It is recorded as typed, never promoted to measured.
• Two analysts disagree. Both saves are in the activity chain with who and when. Do not overwrite quietly — re-save with the rationale that settles it.
• Scoring with the drawer empty. Route the unrouted pile or run discovery first, otherwise every figure on the domain is empty and the score has nothing under it.

Previous: Inside estate import · Next: Rubric

All steps: Analyst method