FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Apex Insurance Brokers
Speak to a brokerGet a quote →

Aipedia: Surveyors and valuers

AVMs and AI valuations: what valuers must do

Valuers can use AVMs and AI pricing tools, but the RICS AI standard requires a written reliability decision on any output that materially affects the valuation. In secured lending, an unchecked model figure is the start of a classic lender overvaluation claim.

In short

An automated valuation model can speed up desktop and portfolio work, but it does not change who is responsible for the figure. Under the RICS AI standard, a named and appropriately qualified surveyor must document whether a material output is reliable (4.2), and high-volume automated work needs randomised dip sampling. The Red Book 2025 adds a commitment to transparency in technology use. If a lender relies on an inflated figure and the loan fails, the claim is usually framed as ordinary valuer negligence, which RICS-compliant PI generally responds to, subject to the policy terms.

How valuers use AVMs and AI pricing

Valuers meet automated valuation models in several ways: as a starting figure for a desktop or drive-by valuation, as a cross-check on a physical inspection, as a screening tool across a lending or investment portfolio, and inside data platforms that suggest a value range from transactions and listings. Newer tools add AI that weights comparables, reads listing text or adjusts for condition. In each case the model produces a number that can travel straight into a report unless someone stops it.

The duties that apply

The RICS AI standard, mandatory since 9 March 2026, treats an AVM output used in a valuation as a material-impact output in most cases. The key duties are:

For Red Book work, the Red Book effective 31 January 2025 includes practice changes for technologies such as AI, with a commitment to transparency in their use and consistent professional judgement, plus new content on modelling, data management and risk. Rule 2 (competence) and Rule 3 (diligent service) of the Rules of Conduct sit underneath. Where an automated tool makes significant decisions about individuals solely by automated processing, the UK GDPR Articles 22A to 22D, in force from 5 February 2026, may also be relevant.

What can go wrong

Illustrative: a valuer produces a batch of desktop valuations for a lender. The model draws on recent sales in a development where incentives inflated headline prices, and the valuer accepts the outputs with light checks. One property later sells well below the reported figure after the borrower defaults, and the lender brings a claim alleging the valuation was outside the acceptable range. The valuer's file shows no written reliability decision and no dip sample for that batch.

The pattern matches the reasons lenders sue surveyors after a fall in the market or a default. The model is not a defence; the question remains whether the figure fell within the range a reasonably competent valuer would have reached. See also the Red Book breach claim scenario.

How PI usually responds

An overvaluation claim arising from a careless reliance on a model is usually treated as a professional negligence claim under the full civil liability cover of the RICS Approved Minimum Wording, subject to the policy terms and the excess. The wording's cyber and data endorsement says cover is not restricted solely because a computer system was used. Losses caused by a computer system failure or unavailability, as opposed to a valuer's judgement, may be treated differently. The limit you need for lending work can exceed the RICS minimums; see Red Book and valuer negligence claims.

Practical controls

Apex arranges PI for surveyors and valuers, including lending panel work. A broker can review whether your limit and wording suit model-assisted valuations. The AI in surveying and valuation hub and the compliance checklist cover the wider duties.

Sources

Frequently asked

Can I sign a valuation based only on an AVM figure?

Not without applying judgement. Under standard 4.2 the output must be assessed with professional scepticism and a written reliability decision made by or under a named, appropriately qualified surveyor. If the output cannot reasonably be used, the client must be told.

Does the Red Book say anything about AI?

RICS says the Red Book effective 31 January 2025 includes practice changes for rapidly developing technologies such as AI, a commitment to transparency in their use and consistent professional judgement, plus confidentiality and data rights protections and new content on modelling and data management.

Will PI cover a claim caused by a faulty AVM?

If the claim alleges the valuer negligently relied on the output, it is usually handled as a negligence claim under RICS-compliant cover, subject to the terms. Losses caused by system failure or unavailability may be excluded by the cyber and data endorsement, so the facts matter.

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.