Seven to eight weeks, timed to include a run of client reviews. Wealth advisers need around 18 hours of learning, two weeks of supervised decisions on live client work, and a final exam and sign-off. The advice stays with the adviser; the training teaches them to test what the agent has assembled before they give it.
What a wealth adviser supervises
Agents in wealth management usually read client files, portfolio data, fact-finds and meeting notes, and draft review packs, suitability reports, rebalancing proposals and follow-up letters. Some, where earned, rebalance model portfolios within tight drift bounds.
The adviser's supervision is about suitability and client circumstances. A rebalancing proposal can be mathematically right and wrong for a client who mentioned at their last meeting that they plan to buy a house. A suitability report can read well and rest on a fact-find two years old. The adviser must spot both.
How the weeks fall
- Week 1: foundations. About four hours: governance, depth, the decision chain, ownership, data and trust.
- Weeks 1 to 5: operator learning. About 14 hours: human review, teaching the agent, keeping knowledge current, when something goes wrong.
- Weeks 5 to 7: supervised client work. Real decisions on held review packs, suitability drafts and proposals, each checked by a senior adviser or the compliance function.
- Week 7 or 8: exam and sign-off.
Advisers have diaries full of client meetings. Book the supervised period across a block of annual reviews so there is enough work to judge.
What the agent prepares and what the adviser keeps
| The agent may | Stays with the wealth adviser |
|---|---|
| Read client files, fact-finds, portfolio data and meeting notes (depth 1) | Any recommendation to a client |
| Draft review packs, suitability reports and letters (depth 2) | Judging whether the fact-find is still current |
| Propose rebalancing within the client's agreed model (depth 2) | Any change to risk profile or objectives |
| Rebalance model portfolios within drift bounds (depth 3, where earned) | Trades outside the model, and withdrawals |
| Hold anything touching a change in circumstances | Conversations about vulnerability, illness or bereavement |
Sources decide most cases. The custodian's position data is sealed. A client's statement of income or plans is declared. A product provider's document is signed. A suitability draft that leans on a declared fact for a decisive point should send the adviser back to the client.
An illustrative example
Take a wealth management firm with ten advisers and an agent preparing annual review packs. Advisers complete Foundations and the Operator modules over five weeks at about three hours a week.
During supervised review, calibration on the same ten review packs, using the method in calibration sessions for wealth advisers, shows that several advisers approved packs recommending no change for clients whose meeting notes mentioned retirement within a year. The agent had flagged the note but not weighed it, because the fact-find still showed a long horizon. The firm adds "retirement date mentioned" to the always-held list. Advisers sit the exam in week eight.
Clients sometimes press for quick trades, and documents can carry instructions aimed at the agent. See training wealth advisers to spot prompt injection and pressure.
How wealth advisers compare with similar roles
HR advisers face similar judgements about people's circumstances, without suitability rules; see how long it takes to train HR advisers. Caseworkers work slower, with welfare stakes; compare the caseworker timeline. A week-by-week plan is in the training plan template for wealth advisers, and the wider approach in training teams for AI agents.
Questions people ask
Can the agent talk to clients directly?
It can draft messages for an adviser to send. Anything that could be read as advice should go out under the adviser's name after review.
What if a client asks whether AI was used?
Tell them, and explain that a named adviser reviewed and stands behind the recommendation. Under the EU AI Act, people must be told when they interact with an AI system.
How often should advisers refresh their training?
Whenever the agent's sources, models or depth change, and at least yearly. Annual reviews are a good moment to run a calibration round.