Client AI disclosure notices: what surveyors must tell clients, and when

The RICS AI standard requires clients to be told about material AI use in writing and in advance. What a client AI notice should say, when to send it, and what else must change in your documents.

Do surveyors have to tell clients they use AI? Under the RICS AI standard: yes — where an AI system has a material impact on delivery of the service, clients must be told in writing and in advance, including when and for what purpose AI is to be used.

That single sentence carries three obligations firms miss: the notice must be written (an email counts; a conversation does not), it must arrive before the AI is used on the instruction, and it must be specific about purpose — “we use AI tools” is not “AI-assisted measurement will be used to extract quantities from your drawings, checked by the responsible surveyor”.

What a good notice contains

• Which parts of the service involve AI — named plainly: measurement extraction, document drafting, transcription.
• Purpose — what the AI contributes to the deliverable.
• Human oversight — who reviews the output and takes responsibility (this is also where your written reliability decisions live).
• The client's routes — how to ask questions, contest AI use, or opt out where offered, consistent with your terms of engagement.

Timing matters more than polish. A short, clear notice attached to the terms of engagement or the instruction-confirmation email satisfies “in advance”; a beautifully drafted letter sent after the cost plan went out does not.

The related duty nobody expects

Section 4.2 adds a mirror-image duty: where a firm decides an AI output cannot reasonably be used for its intended purpose, that conclusion must also be communicated to the client in writing, with the reasoning or a summary of it. Client communication under this standard runs both ways — it is not only “we will use AI”, but occasionally “we used it, and rejected the result”.

Fear of disclosure, and the opposite fear

Firms worry a notice reads as corner-cutting. In practice two pressures are converging: the standard requires disclosure, and clients are increasingly asking the opposite question — whether you are using modern tools efficiently. A clear notice with named human oversight tends to land as professionalism, not confession. The firms that struggle are the ones explaining undisclosed AI use after the fact.

Does every instruction need an AI notice?

Only where AI use has a material impact on delivery of that service. Immaterial back-office use doesn't trigger the notice duty — but the materiality call itself should be recorded per instruction. See how to make and record materiality decisions.

Is a clause in our standard terms enough?

The standard requires both: engagement documents must contain the six AI items (when, which parts, PI cover if available, contest, redress, opt-out), AND clients must be told in writing, in advance, when and for what purpose AI will be used on their matter. A generic clause plus a specific advance notice is the clean pattern.

What if the client objects?

Your terms must set out the internal process to contest AI use and how (if at all) a client can opt out. Decide those answers before the first objection, not during it.

This article is general information, not legal advice — have client-facing wording reviewed for your circumstances. ComplyQS is not affiliated with or endorsed by RICS. Related: the exporting-documents guide at /guides/exporting-documents/.