What counts as Consumer Duty evidence: an audit-ready checklist
Consumer Duty is a proof problem. The evidence has a shape — and an information request wants to see it. These are the artefacts that answer “show me your evidence”.
The evidence is not a single document. It is four connected records a reviewer can follow from method to conclusion.
What the FCA says about Consumer Duty evidence
Under the Consumer Duty, a firm must act to deliver good outcomes for retail customers and be able to demonstrate, openly and honestly, that it is doing so. The FCA's Consumer Duty pages set out that expectation and the supporting rules in full.
Record-keeping failures keep appearing in the FCA's Enforcement Watch newsletter: firms that could not show what happened for their customers because the record was never built.
The four evidence artefacts
Each artefact answers one question a reviewer will ask. Together they form the evidence chain from method to board.
Your file-review methodology
The fixed rubric you score against, how you grade, and how you record results. If a reviewer asks 'show me your methodology', this is it.
Graded logs
The files reviewed, scored Pass / Amber / Fail against the rubric, with the check-area breakdown. The record that monitoring happened.
Overrides and sign-off
Where a human changed an AI or automated score, and why. The audit trail that shows judgment, not just automation.
The board summary
The one-page outcome summary: counts, trends, gaps, fixes. The evidence in the format a board — or a reviewer — can read cold.
None of these need to be elaborate. They need to exist — and to be consistent with each other.
Worked example: the same quarter, with and without the trail
The same quarter looks different with and without a trail. Compare the two versions (anonymised).
Before
“We reviewed 40 files this year. Issues were fixed with advisers.” — no rubric, no log, and no record of what “fixed” meant.
After
“40 files reviewed against revision 4 of the firm's rubric (12 checks). 30 passed; 6 amber on cost disclosure; 4 failed on capacity-for-loss evidence. Overrides logged with reasons. One page to the board with owners attached.”
Common mistakes when building the evidence trail
- A methodology that never stabilises. If the rubric changes between quarters, the graded logs cannot be compared and the trend line means nothing.
- Logs without sign-off or overrides. A score with no human review and no record of the judgment reads as automation output rather than evidence.
- A board summary that reports intentions. Counts of files reviewed are not the same as outcomes demonstrated; a reviewer looks for the fixes and the owners.
- Artefacts kept in separate places. A rubric in one system, logs in another, a board summary nowhere — the chain breaks. the smallest defensible process is the lean version built to survive a cold read.
The evidence checklist
| Artefact | What it must contain | Why a reviewer reads it |
|---|---|---|
| Methodology | The versioned rubric and grading rules. | To check the same checks applied to every file. |
| Graded logs | Files scored Pass / Amber / Fail with dates. | To see monitoring happened, and when. |
| Overrides and sign-off | Who changed a score, and why. | To see judgment, not just automation. |
| Board summary | Counts, trends, gaps, and fixes. | To see the loop closed in front of the board. |
Related reads
This guide sits inside the outcomes monitoring cycle — start at the outcomes monitoring hub , then the in-practice examples for what good firms actually track, and the four numbers every board report relies on to make the MI visible.
Frequently asked questions
Do these artefacts need to be formal documents?
No. The Consumer Duty asks a firm to demonstrate good outcomes; the artefacts can be simple, but they must exist, be consistent with each other, and be findable. A reviewer reads what is in the file, not what was intended.
How do the artefacts differ from the annual assessment?
The artefacts are the ongoing record each quarter produces: methodology, graded logs, overrides, and the board summary. The annual assessment is the yearly review of that whole cycle. An information request tests both.
What if one artefact is missing?
The artefacts work as a chain, so a missing link weakens the rest. A firm that cannot produce its methodology or its graded logs has to rebuild the record at the moment it is asked for — the slowest point at which to do so.
Your next step
Take the quarter you already ran and shape it: version the rubric, keep the graded logs, add an override line, and put one page in front of the board. Proven Duty produces the graded log and the board summary from the same review. Start with a free trial or see pricing.
Not sure where your firm sits? Run the 4-minute evidence scorecard and get a grade with the fixes.