Aipedia: Financial advisers

AI in cashflow modelling and planning

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

AI can speed up lifetime cashflow and drawdown modelling, but the adviser remains responsible for the assumptions and for whether the resulting retirement advice is suitable. A flawed model can make an unsustainable income plan look safe, and the client may not find out for years.

Part of: AI in Financial advisers

In short

Cashflow and drawdown tools increasingly use AI to suggest assumptions, run scenarios and explain results in plain English. The risk is that growth, inflation, charges or withdrawal assumptions go unchallenged because the output looks authoritative. COBS 9.2.1R requires the recommendation to be suitable, the Consumer Duty requires firms to avoid foreseeable harm, and COCON Rule 2 requires due skill, care and diligence. Claims tend to arise late, when a client's fund runs short. PI usually treats these as negligent advice claims, subject to the wording.

How advisers use AI in cashflow modelling

Lifetime cashflow modelling sits at the heart of retirement and later-life planning. AI features now pre-fill income and expenditure from fact-finds, suggest assumptions, run stochastic or scenario analysis, recommend sustainable withdrawal levels and write the narrative that goes into the client report. Some tools offer conversational interfaces that let advisers ask "what if" questions in plain language.

The model often becomes the centrepiece of the client conversation. Clients remember the chart showing their money lasting to a given age far more than the caveats around it. That makes the model, and the assumptions behind it, part of how the advice is understood, which brings the Consumer Duty consumer understanding outcome into play as well as suitability.

The duties that apply

Suitability: COBS 9.2.1R(1)(a) requires reasonable steps to ensure the recommendation is suitable. A drawdown plan built on an unrealistic model may fail that test.
Consumer Duty: PRIN 2A.2.8R (avoid causing foreseeable harm) and the consumer understanding outcome, which bears on how model limitations are explained.
Individual conduct: COCON Rule 2 (due skill, care and diligence) and Rule 6 (good outcomes for retail customers).
Systems and controls: the FCA AI Update maps safety and robustness to Principles 2 and 3, SYSC 4 and SYSC 7, and third-party tools to SYSC 8 outsourcing.

What can go wrong

The main risks are assumption errors: growth rates that do not reflect the actual portfolio, inflation set too low, charges omitted, state pension timing wrong or life expectancy understated. AI can also generate a narrative that overstates certainty, presenting a projection as if it were a plan the client can rely on. Where a tool suggests a "sustainable" withdrawal rate, advisers may accept it without testing it against the client's real portfolio and capacity for loss.

Illustrative scenario: a firm adopts an AI feature that sets default assumptions. It applies a growth rate suited to a higher-risk portfolio to clients in a cautious model portfolio. Several drawdown clients are told their income is sustainable to age 95. A file review years later finds the models were never reconciled with the actual investments, and a number of clients face cutting their income.

How PI usually responds

Modelling-based complaints usually present as unsuitable retirement or drawdown advice, so PI commonly responds as it would to other advice claims, subject to the policy terms and excess. Because a default assumption can affect many clients, insurers may look at aggregation and at when you first became aware of a problem. Late notification can cause difficulties, so treat a systemic modelling issue as a potential circumstance. See a pension freedoms PI scenario and how PI responds to suitability claims.

Practical controls

Set firm-wide assumption policies and record who approved them; do not rely on vendor defaults.
Reconcile model assumptions with the client's actual or recommended portfolio and charges.
Stress-test outcomes for poor sequences of returns and higher inflation, and show clients the downside.
Edit AI-written narratives so they describe projections, not promises.
Re-run the model at each review and record why assumptions changed.
Link the model to the attitude to risk and capacity for loss assessment.
Keep model versions and inputs with the advice file and review them at each file check.

How Apex helps

Apex arranges PI for financial advisers. A broker can review your cover and how your policy treats AI-assisted work, including retirement income and drawdown advice. See AI in financial advice and our guide to AI in pension transfer analysis.

Sources

Frequently asked

Is the adviser responsible for assumptions the AI tool suggests?

Yes. COBS 9.2.1R requires the firm to take reasonable steps to ensure the recommendation is suitable, and SYSC 8.1.6R keeps the firm fully responsible for outsourced functions. Accepting vendor defaults without review is unlikely to meet COCON Rule 2.

Can a cashflow model be shown to clients as a forecast?

It can be shared, but the Consumer Duty consumer understanding outcome means clients should understand it is a projection with limits. Show downside scenarios and avoid AI-generated wording that suggests certainty about future income.

How do PI insurers view a systemic modelling error?

Usually as a potential series of related advice claims. How the policy aggregates them, and when you notify, can affect how the limit and excess apply. Review the aggregation and notification terms with your broker.

Should I review cashflow models at each ongoing review?

It is good practice. Markets, charges, spending and health change, and a model that was reasonable at outset may no longer support the income being taken. Recording why assumptions changed also gives you evidence of ongoing oversight if a client later complains that their income was unsustainable.

Check your PI covers how you use AI

Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start a PI quote Request a callback

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.