Skip to content
Keel Docs

Evidence-readiness meter

The evidence-readiness meter answers a plain question: how audit-ready is your evidence, really? It reads as “N of M expected evidence artifacts present (P%)” and is derived entirely from evidence you already have — there is nothing extra to fill in.

What it measures

Section titled “What it measures”

The meter compares two numbers:

  • M — expected artifacts (the denominator). For each control in your program, Keel’s “evidence to collect” guidance already lists the concrete artifacts an auditor typically expects for that control (a policy document, an access-request approval, a review sign-off, and so on). The number expected of a control is exactly the length of that guidance list. Summed across the in-program controls that apply, that is your expected total. It comes from Keel’s authored guidance, so it is never an invented or inflated figure.
  • N — artifacts present (the numerator). This is how many evidence artifacts you’ve linked to those controls. It’s counted per control and capped at what the control expects, so the meter can never over-state your readiness: five files linked to a control that expects three still reads as “3 of 3”, never “5 of 3”. This conservative count is the honest one — Keel can’t map an individual file to a specific expected idea, so it never claims more coverage than the guidance says a control needs.

P% is simply N / M as a percentage. When nothing is expected yet — no in-program controls have authored guidance — the meter shows a gentle empty state prompting you to apply a framework and start linking evidence.

Why the denominator is scoped to your program

Section titled “Why the denominator is scoped to your program”

The meter is deliberately about evidence, not control coverage. It only counts controls that are already in your program. A control you haven’t adopted yet is a controls-coverage gap — that’s what your readiness score reflects — not an evidence gap. Keeping the two separate means the evidence-readiness meter tells you specifically whether the controls you’ve committed to have the proof an auditor will ask for.

Where it appears

Section titled “Where it appears”
  • Dashboard. The meter reflects the framework you’re currently viewing and follows the framework switcher, so you can see how audit-ready the evidence is for one framework at a time. When a framework has no in-program controls with guidance, it falls back to your whole program.
  • Questionnaire assist. Here the meter covers your whole program, so before you send answers grounded in your controls and policies you can gauge whether the evidence actually backs them up.

How it relates to the AI evidence review

Section titled “How it relates to the AI evidence review”

The meter and the AI evidence review answer different questions. The meter is a fast, credit-free count of how much expected evidence is present across your controls. The AI evidence review is an optional, credit-metered check of whether the specific artifacts on one control are strong enough — it weighs each item’s label, kind, and freshness and returns a hedged verdict. Use the meter to see where to look; use the AI review to judge a control’s evidence in depth.

Next steps

Section titled “Next steps”