A Consumer Duty board report where every number is cited to its data and every weak point is challenged before the board sees it
Every board report sentence tied to its data, and every weak point challenged first.
See one case, screen by screen ↓Consumer Duty Outcomes Monitoring helps a UK bank's conduct team monitor Consumer Duty outcomes and fair value and draft the Consumer Duty board report. Agents assemble the outcome MI from the bank's own systems and write each finding as one sentence whose numbers are live and cited, and a Challenge agent questions the draft the way the FCA reviewed board reports and cash savings. The Conduct Risk Lead answers, product owners commit costed actions, a pricing approver signs any rate change, and the Consumer Duty Champion approves the report.
AI use cases in banking: a buyer's guide to agentic AI for bank and credit union operations →
Updated
A Consumer Duty board report has to tell the board, at least once a year, what outcomes retail customers actually get under PRIN 2A: products that meet their needs, fair value, communications they understand and support that works, including for customers in vulnerable circumstances. The Conduct Risk Lead assembles it by hand, requesting MI from six teams and about twenty sources, then writes findings such as “customers on closed savings accounts receive fair value” from memory and a complaints count. Since July 31, 2024 that includes every closed product.
The FCA read 180 firms' first board reports and found thresholds set with no reasoned justification, claims that did not seem to align with the MI provided, vulnerability treated as a catch-all category, and actions that lacked owners and timescales. Its 2024 review of the nine largest easy-access savings providers found fair value assessments that validated old pricing rather than testing it, and named telling customers they are on a low rate, rather than addressing the harm, as poor practice. Built by hand, the report takes weeks, which leaves little time for the board's own challenge.
Estimated split for one annual report at a mid-size UK retail bank, by hand and with the solution; the 2-day total with the solution is its target.
Seven agents assemble the MI, draft cited sentences and challenge the draft; the Conduct Risk Lead answers, product owners commit, and the Champion approves.
Corvane Bank UK has 1,214,600 customers and 46 retail products, 19 of them closed savings issues. Its Consumer Duty annual board report 2026 covers October 1, 2025 to September 30, 2026; the board meets on November 26 and the board pack closes on November 12. Here is one morning's work on it, screen by screen, in the working solution.
Imogen opens draft v3: four sections, one per Consumer Duty outcome, holding 14 one-sentence findings with 32 live numbers. Two findings carry “No evidence yet” and one carries “Data late”, because the contact center transcripts last loaded on October 19. The right column lists the challenges already answered on the draft, and who answered each.
“Before it can go to the Champion: The challenge review has not run on draft v3.”
Sentence 2.1 reads: Easy Access Saver Issue 3, 41,800 customers earn 1.10% AER while the current issue pays 3.85%; balances of £27.4 million; 62% of holders are aged 70 or over. Opening it shows each of its five numbers with its system, query, as-at time and confidence: 41,800 from the core banking system, query “Open accounts, product EAS-I3”, as at Sep 30, 2026 · 23:59, confidence 0.99. The two rates come from the rates history.
Below a confidence of 0.85, a person confirms each number before the sentence can be used.
Imogen runs the challenge review. The Challenge agent reads the draft against the FCA's published findings on board reports and on cash savings, and raises CH-1 on sentence 2.1, due October 29. Issue 3 is the largest poor-value issue and no action covers it.
“Were these 41,800 customers told that the current issue pays 3.85%?”
Imogen asks the agents to check. The Communications Trace agent searches letters, app messages and call transcripts from October 2022 to October 2026 for the 41,800 holders: 2 reminder letters, the savings summary of March 14, 2023 and the annual statement of February 10, 2025; no app message, with 71% of holders not logged in to online banking in 12 months; and 204 of 312 calls about their rate where the adviser did not mention a better account.
The trace agent hands its findings back; only a person answers a challenge.
Imogen answers CH-1 with action A-07, “Move Issue 3 holders to the current rate, or raise Issue 3 to it”, owner Rohan Mehta, due January 31, 2027, reason code CH-A3. The sentence now carries the action, and every challenge on v3 has a recorded answer, so the draft can go to the Champion.
“CH-1 answered · every challenge on v3 now has an answer.”
The fair value assessment for Issue 3, closed to new customers in September 2019, draws its rate flat near 1.10% against the issue on sale rising to 3.85%, with the two letters marked on the line. It shows who holds it: £27.4 million, a median balance of £650, and 6,480 holders on the vulnerability register. The assessment has a maker and a checker: Rohan signs it, and Imogen checks it after him.
The Fair Value agent costs three options: raise Issue 3 to 3.85% AER from December 1, 2026, an estimated £753,500 a year; move holders into Issue 7, which costs the same and needs 41,800 notices; or write to holders only, marked not enough alone. Rohan, the Savings Product Owner, commits the rate rise. Callum Brereton, chair of the Retail Pricing Committee, approves it with reason code PR-01, and the customer notice is drafted at reading age 9.
The proposer can never approve his own rate change.
Imogen submits v4. Margaret, an independent non-executive director, sees what changed since v3, that 6 of 6 challenges are answered and that 2 gaps are reported to the board with actions A-09 and A-10. She approves v4 for the November 26 board with reason code RP-01, and the Evidence Pack agent seals it and publishes it to the board reporting portal.
Imogen cannot approve her own draft; Margaret cannot edit a sentence.
The audit trail holds every step, newest first: the challenge, the trace with its letter references and redacted transcripts, the action, Rohan's commitment, Callum's approval with pricing paper PP-26-131, the submission and the Champion's approval. It exports as a CSV or as sealed evidence pack EP-2026-CD-04 for the FCA or internal audit.
The same report, seen by four of the people who carry it: what their part looked like, and what it looks like now.
Seven agents assemble the MI, draft cited sentences and challenge the draft; the Conduct Risk Lead answers, product owners commit, and the Champion approves.
Assembles the outcome measures per product, outcome and cohort from the systems of record, records the query and as-at time with every value, and watches how fresh each feed is.
Writes each finding as one plain sentence with its numbers bound to the MI agent's queries, marks a finding it cannot support “No evidence yet”, and drafts customer notices.
Assesses price against benefits for every product, compares each closed issue with the issue on sale, and costs each option for the product owner.
Tests “were they told?” across letters, app messages and call transcripts for the holders of a product.
Compares outcomes by FG21/1 driver (health, life events, resilience and capability) with all customers, and notes where a cohort fares worse.
Reviews the draft against the FCA's published good and poor practice on board reports and cash savings, raises one question per weak point, and holds the draft back until each is answered.
Seals each approved version with its evidence, publishes it to the board reporting portal, and builds export packs for the FCA and internal audit.
Answers every challenge; the Champion approves the report. The agents propose; a named person decides.
Ask about any sentence, number, product, cohort or challenge in plain words. Every answer points to its source.
Which closed issues pay more than 2 points below their current issue?
Five, holding £61.9 million for 88,300 customers: Postal Saver (2.95 points below, 74% aged 70 or over), Easy Access Saver Issue 3 (2.75), Issue 1 (2.65), Bonus Saver Issue 1 (2.35) and Instant ISA Issue 1 (2.25). Postal Saver is covered by A-04 (approved) and three by A-05; Issue 3 has no action yet.
Were Issue 3 holders told about the better rate?
Twice in four years: the savings summary of Mar 14, 2023 and the annual statement of Feb 10, 2025. No app message reached them, and 71% had no online banking login in 12 months. In 204 of 312 calls about their rate, the adviser did not mention a better account.
What is still stopping the report going to the Champion?
Two things decide it: the challenge review must run on draft v3, and every challenge must be answered. The two “No evidence yet” sentences do not block it: each is reported to the board as a gap with an action (A-09, A-10).
How do customers with low capability fare against everyone else?
Worse on price and on support: 16.5% hold a poor-value closed issue against 7.3% of all customers, and they wait 9 min 30 s for the contact center against 6 min 40 s. Annex A has all four drivers; one sub-cohort is too small to report.
13 screens from the working solution, on its sample data. Pick one to see it large.
Every outcome threshold carries the written reason behind it, a change needs a second person, and every setting is on the audit trail.
| Setting | Default | Choose from |
|---|---|---|
| Rate gap between a closed issue and the issue on sale | 2.00 points | Reason required (Retail Pricing Committee minute PC-25-41); a change waits for a second person |
| Complaints with a final response within eight weeks | 95% | DISP 1.6.2 sets eight weeks; the floor is the Conduct Risk Committee's |
| Minimum cohort to report | 50 customers | A smaller cohort shows “Too small to report” |
| Confidence threshold | 0.85 | Below it, a person confirms each number before the sentence is used |
| Challenge answer | 5 business days | From the day the challenge is raised; the board pack closes 10 business days before the board |
| Action reminders | 10 days before due | Overdue actions escalate to the Consumer Duty Champion |
| Human override | Reason code OV-01 required | A person may correct any number or sentence |
| Retention | 6 years | Versions, evidence and the trail, under the bank's records policy |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (FCA, Consumer Duty board reports: good practice and areas for improvement, December 11, 2024, updated March 9, 2026: 180 firms' first board reports reviewed; FCA, Update on cash savings, September 18, 2024: the nine largest easy-access savings providers reviewed). Rules named: FCA Handbook PRIN 2A, FG21/1 and DISP 1.6.2. Amounts are in pounds sterling because the sample bank reports in pounds. Corvane Bank UK and the people, products and references named on this page are fictional: characters and sample data in the working solution.
Consumer Duty Outcomes Monitoring helps a UK bank's conduct team monitor Consumer Duty outcomes and fair value and draft the Consumer Duty board report. Agents assemble the outcome MI from the bank's own systems and write each finding as one sentence whose numbers are live and cited, and a Challenge agent questions the draft the way the FCA reviewed board reports and cash savings. The Conduct Risk Lead answers, product owners commit costed actions, a pricing approver signs any rate change, and the Consumer Duty Champion approves the report.
Under PRIN 2A.8, at least once a year the board must receive a report on the outcomes the firm's retail customers receive and the actions taken, and confirm whether it is satisfied the firm complies with the Duty. In the working solution the report opens with that question, then gives one section per outcome: products and services, price and value, consumer understanding and consumer support. Every finding cites its data, and every action has an owner and a date.
The FCA reviewed 180 firms' first board reports and published good practice and areas for improvement in December 2024, updated in March 2026. It found thresholds with no reasoned justification, claims that did not seem to align with the MI, vulnerability treated as a catch-all category and actions without owners or timescales. The Challenge agent tests a draft against those findings, so each threshold here carries a written reason and each action an owner and a date.
Closed products came under the Duty on July 31, 2024. The Fair Value agent compares each closed issue's rate with the issue on sale, shows who holds it and what they were told, and costs each option, such as raising the rate or moving holders. The product owner signs the assessment and commits an action, conduct risk checks it, and the chair of the Retail Pricing Committee approves any rate change.
The Vulnerability Outcomes agent compares outcomes by the four FG21/1 drivers (health, life events, resilience and capability) with all customers, instead of one catch-all figure. Any cohort under 50 customers shows “Too small to report”, and no customer-level record appears in the annex. The Head of Customer Vulnerability signs Annex A.
The Drafting agent never writes a number it cannot cite. A finding with no evidence carries “No evidence yet” and goes to the board as a gap with an owned, dated action; a finding that relies on a late data feed carries “Data late” until the feed loads.
Yes. Agents assemble, draft, trace and challenge; people decide. Only a person answers a challenge, commits an action, approves a rate or approves the report, each with a reason code, and no one approves their own work. Any number or sentence can be corrected by a person with a reason, and the correction is on the audit trail.
The Agentic Solution Engine builds and deploys it from your requirements and documents: your outcome measures and thresholds, your products and your report template, connected to your core banking system, rates history, complaint management, contact center transcripts and vulnerability register. It goes live once every quality gate has passed.
Cited answers while the member is on the line, and Reg E claims opened on the first call.
1-min demoFDIC Signage ComplianceEvery screen checked against Part 328 monthly; no finding closed without a recapture.
1-min demoOverdraft ComplianceEvery overdraft and NSF fee tested, not a sample, and refunds sized before the exam.
1-min demoWe'll run Consumer Duty Outcomes Monitoring on a sample of your own outcome MI and products.