Ongoing Advice Evidence for UK Wealth Firms in 2026

A wealth or advice firm that charges an ongoing fee has always had to deliver the service behind it. What has changed is how hard the regulator now looks at the proof. The FCA's multi-firm review of ongoing financial advice services, published on 24 February 2025, asked 22 of the largest advice firms for seven years of delivery data, and its next steps asked every firm to consider whether it can evidence that it delivered what its contracts required.

The next shift is already on the table. In CP26/10, published on 25 March 2026, the FCA proposes replacing the annual suitability review with periodic reviews set by client need. That sounds like a lighter load. In practice it moves the burden from a calendar to an argument: a firm that reviews some clients every two years and others every year has to show, client by client, why each frequency is fair value for each fee. About 80% of revenue from adviser charges comes from ongoing services, according to the same paper, so this is the centre of the advice business, not a corner of it.

This guide is written for the COO and the Head of Regulatory Affairs at a UK wealth or advice firm who own that argument. It covers what the FCA has said good evidence looks like, what CP26/10 proposes, how to handle clients who have stopped engaging, and what a client-level evidence ledger contains. It deliberately cedes two things to our wealth, PE and hedge fund industry guide: the headline findings of the 2025 review and the history of one large firm's remediation provision. Both are covered there in full. More of our sector material sits under industry guides.

This guide is shipped by the team behind DATS, the five-stage AI consulting system DILR.AI runs for enterprises. Or see the AI operating model, which covers governance, RACI and lifecycle, audit-ready by design.

What does ongoing advice evidence mean for a UK wealth firm in 2026?

Ongoing advice evidence is the record that proves a UK wealth or advice firm delivered the ongoing service a client paid for: the agreement that defined it, the review that was offered, what happened next, and the fee decision that followed. Because the FCA's view on redress turns on what happened in each relationship, the record has to work client by client, not as a firm-wide percentage.

The distinction matters because an aggregate number hides exactly the cases a regulator cares about. A firm can report that most reviews were delivered and still owe redress to the clients in the remainder. The 2025 review drew that line itself: where a firm made no effort to deliver a review, the FCA said redress is likely to be due, and where a firm was ready, willing and able but the client declined, redress is less likely. Those two outcomes can only be told apart if the record for each client shows which one happened.

The duty here binds the advice firm. The Consumer Duty, in force since July 2023, applies to the firm's products and services, and the price and value outcome sits in PRIN 2A.4. A firm that has already worked through Consumer Duty barriers on the phone channel will recognise the pattern: the regulator asks for outcome evidence, not process descriptions. That is the kind of evidence problem DATS consulting is built to place AI against, and also where it should not.

What did the FCA say good ongoing advice evidence looks like?

The FCA's February 2025 review of ongoing advice named four good practices: client agreements that clearly set out the nature and timing of the service, effective systems and resources to schedule and offer reviews as agreed, policies to stop collecting fees where a client had not engaged for a period, and appropriate record-keeping so firms can evidence delivery. It also named four poor practices.

Read together, the two lists describe one system with five parts. The FCA did not pair them one to one; the table places each where it bears. Every part produces a record, and a gap in any one of them breaks the chain.

The review also listed what a good suitability review covered when it did happen: up-to-date circumstances, objectives, attitude to risk and capacity for loss; a review of the risk profile, charges and performance of existing investments; and communications that recorded the outcome as a personal recommendation. Our guide to mapping AI risk to the FCA rulebook uses the same habit of turning a regulator's list into owned controls.

The FCA put one line in the good practice list that deserves a wall of its own in any compliance team: "Appropriate record-keeping to make sure firms can evidence delivery of their services where provided." The rest of this guide is about building that record in a way that survives a change in the review rules. That is the core of our approach to placing AI inside regulated workflows.

What does CP26/10 propose for annual suitability reviews?

CP26/10 is an FCA consultation, published on 25 March 2026 and closed on 22 May 2026, that proposes removing the annual suitability requirement in COBS 9A.3.9R and 9A.3.10R for MiFID business and insurance-based investments. Firms providing an ongoing advice service would instead conduct periodic suitability assessments at a frequency set by each client's needs and circumstances, in keeping with the Consumer Duty.

It is a proposal, not a rule. The FCA said in paragraph 1.25 of the paper that it would aim to publish a Policy Statement in Q4 2026, depending on the volume of feedback, and at the time of writing we could not locate one. Until final rules are made, the annual requirement still applies where COBS 9A applies today. Any firm redesigning its service now should design for both states: annual where the current rules bite, periodic where final rules allow it.

The paper explains the reasoning. Consumers have different life stages and risk tolerances; annual reviews are resource intensive and push up charges; and the Consumer Duty already requires firms to evidence that the services they charge for provide fair value, which the FCA says reduces the need for a prescriptive minimum frequency. It also proposes new rules in COBS 9C.4 that would require firms to make clear to clients the scope and nature of the ongoing service, including how often suitability will be reviewed, and to apply that to all investment advice in scope rather than only MiFID business.

Two further proposals touch charging. The FCA proposes to clarify that firms can charge for related services linked to an earlier personal recommendation, even without a new recommendation each time, and that ongoing payments for an initial recommendation should only be taken until the up-front cost has been paid off. Firms that already record calls for investment business, covered in our note on MiFID call recording for investment firms, will find some of the delivery evidence they need already exists. Using that existing data before building anything new is usually the cheaper first step.

How should a firm evidence ongoing advice when reviews become periodic?

A firm evidencing ongoing advice under a periodic model should, in our view, record three things per client: the review frequency it chose, the reason that frequency fits the client's needs and circumstances, and the link between that frequency and the fee. The FCA's own wording is that a firm must be able to show the service provided is commensurate with the fee charged, which is why we treat the frequency rationale as part of the evidence.

Under an annual rule, the frequency needed no justification: the rule set it. Under a periodic model, the frequency becomes a decision, and a decision that sets the level of service against a fee is exactly what the price and value outcome tests. A client reviewed every two years at the same fee as a client reviewed every year is not automatically a problem, but the firm has to be able to say why, using the client's needs and circumstances, the basis CP26/10 proposes for setting frequency.

Practically, that means adding fields that most back-office records do not carry today: a segment or service tier, a documented frequency, the trigger events that would bring a review forward (a large withdrawal, retirement, a bereavement, a change in attitude to risk), and a periodic check that the frequency is still right. CP26/10 also notes that the Handbook does not prescribe a medium for the review, so a firm can mix meetings, calls and digital fact-finds; recording which was used and what it produced keeps the evidence complete.

Building controls like these inside the business is the work of an AI execution office, embedded delivery where the production placements stay with the client. A board that wants to decide where this kind of work should sit in its investment plan can use our guide to board-level AI investment decisions.

There is a quieter risk in the proposal, and the FCA names it. In CP26/10 it asks whether its rules, the Consumer Duty and supervision would be enough to stop firms reducing service quality without a matching reduction in price. A firm that moves clients to less frequent reviews without moving the fee, and without a recorded rationale, is the case that question is written about.

What should a firm do about disengaged clients who are still paying?

A wealth or advice firm with clients who pay an ongoing fee but no longer engage needs a written policy for when disengagement means the client is not receiving the service, and that policy should include stopping fee collection. The FCA says there are no specific Handbook rules on disengaged clients today, so the expectation comes from the Consumer Duty, and CP26/10 proposes guidance rather than new rules.

The FCA's position in CP26/10 is direct. It found in 2025 that many firms already had policies to stop collecting fees after a period of non-engagement, but that practice varied widely, which raised concerns that some firms may not be meeting their Duty obligations. It considered prescriptive rules on when charging should cease and said it does not think they are the right approach, because disengagement can be temporary and does not always mean a client no longer benefits from the service. Then it added one sentence that every fee policy should be tested against: "We expect firms are mindful that continued charging without meaningful delivery would ultimately breach their Consumer Duty obligations." (FCA, CP26/10, March 2026.)

The paper also invites examples of good and poor practice in six areas, and they double as a design brief for the policy:

• how the firm defines a disengaged client, including timeframes and behavioural markers;
• what systems identify clients who are not engaging, and how the firm assesses whether they still get value;
• how the firm tries to re-engage clients who have gone quiet;
• what happens once every attempt to make contact has been exhausted;
• how the firm decides whether fees continue, stop or are refunded;
• how engagement and disengagement activity is recorded so the record is complete, accurate and up to date.

The 2025 review adds the remedy view. Where a client has declined the service for several years, firms should discuss whether continuing is still in the client's best interest. Where a firm made reasonable and proportionate attempts to engage without success, redress is less likely, and a firm can only show those attempts were made if they are on the record. If your team wants to talk through a fee-stop policy against your own data, talk to the DATS team.

What does a client-by-client ongoing advice evidence ledger contain?

A client-by-client ongoing advice evidence ledger is one record per client relationship that links the agreement, every review offer, every delivery or non-delivery outcome, every engagement attempt and every fee decision, with dates and owners. It lets a UK wealth firm answer a regulator, an ombudsman or its own board about any single client without rebuilding the history from emails and diaries.

The chain below is the minimum. Each node produces a record the next node depends on, and the fee decision at the end is only defensible if the four before it exist.

In data terms, that is a small number of joined tables: the client agreement and its service schedule, a review events table, an engagement attempts table, and a fee decisions table, all keyed to the client and the adviser. Most firms hold the raw material across a back-office platform, a document store, the adviser's calendar and the platform that collects charges. The work is the join, and the exception report that falls out of it: every client being charged whose last delivered review is older than their recorded frequency allows.

That exception report is where AI earns a place, and it sits close to the enterprise knowledge retrieval and evaluation work listed on our enterprise AI solutions page. The same pattern of reconciling regulated records at client level drives our analysis of KYC review cost in UK banks. For a firm whose ledger sits on Azure, AWS, Snowflake or Databricks, the architecture choice matters less than the ownership of each table.

How far back should an ongoing advice remediation look, and who owns it?

The FCA said firms carrying out proactive reviews of ongoing advice delivery should look back to 2018, because its record-keeping rules and the MiFID II and insurance-based investment requirements for annual suitability reviews took effect then. Ownership sits with the advice firm, and in our experience a remediation moves faster with one named senior owner, recorded in the firm's Senior Managers and Certification Regime responsibilities map where it has one.

The 2025 review was specific on the rule set. Under the Consumer Duty and the complaints sourcebook, firms must take appropriate action where retail customers have suffered foreseeable harm, and it pointed to DISP 1.3.6G and PRIN 2A.2.5R when considering redress. It also allowed that redress might cover only part of the charges where some services were delivered, which is fact dependent. Partial redress is only arguable with partial evidence: a firm that cannot show which years were delivered cannot argue for a part refund.

In CP26/10 the FCA says it is now asking a small number of firms to outline the actions they have taken since the review, including remedial steps. For a COO, that makes the look-back a current supervisory question, not an archive project. The practical sequence is to build the ledger for live clients first, run the exception report, and then extend the same structure backwards year by year to 2018, so the historic work uses the same definitions as the forward controls.

Ownership is where many programmes stall. A remediation that sits with compliance but needs data from operations and changes to charging in finance needs decision rights written down before it starts, which is the point of our comparison of an AI operating model versus a centre of excellence.

Where can AI help with ongoing advice evidence, and where must a person decide?

AI can help a UK wealth firm assemble ongoing advice evidence: matching review records to clients, reading meeting notes and suitability letters to classify outcomes, flagging clients whose fees continue past their review frequency, and drafting re-engagement messages for an adviser to approve. It should not make the personal recommendation, set a client's review frequency, or decide to stop, continue or refund a fee.

The FCA has left the door open. CP26/10 notes that the Handbook does not prescribe a medium for the suitability review, says that if the annual requirement is removed firms should be able to use technology both to assess suitability on an ongoing basis and to deliver the review, and observes that the role of AI is becoming increasingly prominent. Our recommendation is narrower: automate the evidence work first, and keep the personal recommendation with the adviser and the fee decision with the firm, which owns it under the Consumer Duty.

In practice the useful placements are narrow and checkable. Classification of historic review outcomes from documents is a good first candidate, because a sample can be checked by a compliance analyst and the error rate measured before scale. Exception reporting on fees against frequency is a second, because it is a deterministic join with a small model-assisted step for messy records. Evaluation and observability for production agents, one of the solutions DILR.AI lists, is the relevant discipline here: measuring the error rate before a regulator asks.

Two neighbours are worth reading before scoping. The FCA AI governance note sets out how the regulator frames AI use under existing rules, and the guide to building an AI tool inventory for the FCA and ICO covers the register any new placement should join. Firms using outbound calls to re-engage clients can look at Dilr Voice for the calling layer, with the adviser still owning the conversation that matters.

What is the best way to build ongoing advice evidence in 2026?

The best way for a UK wealth or advice firm to build ongoing advice evidence in 2026 is to design one client-level ledger that works under both the current annual rule and a periodic model, run a fee-against-frequency exception report on live clients first, and only then extend the same structure back to 2018. The right partner depends on whether the gap is data, process or a regulatory interpretation.

Each route fits a different situation. In our assessment, a large consultancy such as Accenture, Deloitte, PwC or EY can bring scale, and for a firm facing a formal supervisory request across many thousands of clients with a fixed deadline, that capacity can be the deciding factor. Data specialists such as Faculty are a natural fit when the problem is mainly a modelling one. The back-office platform vendor may simply be the fastest route where the gap is a missing field or report, and a firm should check that before commissioning anything larger.

DATS fits the case in between: a firm that wants an operating model naming who owns frequency decisions and fee stops, and embedded delivery on its own data by senior practitioners who ship code, not decks. The diagnostic produces a ranked roadmap of where AI belongs in this work and where it does not, which is often the more useful half. Our comparison of an embedded AI delivery team versus project consulting sets out the trade in more detail, and the full method is on the DATS services page.

The concession is real. If the firm's gap is a single missing report in an otherwise sound back office, the platform vendor or an internal analyst will close it faster and more cheaply than any consultancy, DATS included. Firms running agent-based operations desks can also look at Cognibl, from DILR.AI, for governed task work with proof of completion, though the advice judgement stays with the adviser either way.

Frequently asked questions

Does the annual suitability review rule still apply today?

The annual suitability review rule under COBS 9A.3.9R and 9A.3.10R still applies to UK firms giving ongoing advice on MiFID business and insurance-based investments. CP26/10 only proposes replacing it with periodic reviews. The FCA said it aimed to publish a Policy Statement in Q4 2026, and until final rules are made and in force, firms should keep delivering and evidencing annual reviews where the current rule bites.

That is also why the ledger design above records a frequency rationale today: under the annual rule the rationale is simply the rule, and the same field carries a client-specific reason once a periodic model is allowed. The wider sector picture is in our wealth and private markets AI guide, linked in the introduction.

Can a firm keep charging a client who never responds to review offers?

A UK advice firm should keep charging a non-responsive client only where it can show the client still benefits from the ongoing service, and the FCA accepts disengagement can be temporary. The FCA expects firms to have a policy for disengaged clients that includes stopping fee collection, and it has said continued charging without meaningful delivery would ultimately breach Consumer Duty obligations. Each attempt to re-engage should be recorded with dates.

Teams that need to train advisers and paraplanners on the new evidence habits can look at Dilr Academy for institution-built courses alongside the process work.

Want to see this in production? Read how DATS consulting places AI in regulated firms, see the embedded execution office, read the UK AI consulting guide, or read about our approach to placing AI inside enterprise systems.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed.