The AI gap now landing on surveyors’ desks

New: the full AI in Surveying briefing now comes by email, every week, free. Join here.

A firm told me recently about a client who was making a complaint.

The client would only communicate through AI.

Your clients adopted AI faster than your firm did. That gap is your exposure.

In February I wrote that your report is already inside ChatGPT [1]. Clients were uploading surveyors’ reports and getting a list of challenges and a draft complaint back in seconds.

Nine months on, it isn’t just reports. AI is increasing everything: queries, challenges to quotes and instructions, complaints and claims.

Clients moved first

More than half of UK adults (54%) now use AI tools such as ChatGPT, Copilot or Gemini, rising to 79% of 16 to 24-year-olds [2].

Firms are further behind. RICS’s 2026 AI report, drawing on more than 3,100 surveyors and other professionals, found that the most common stage is still an early pilot, and that embedded use is rare [3]. The RICS AI standard itself admits that members’ knowledge of AI is “uneven” [4].

So on one side, clients use AI every day, with no policy, no training and no rules. On the other, firms are still working out what their policy should say.

That’s the gap.

What’s landing on your desk

Whether it’s a homebuyer’s survey, a dilapidations schedule, a valuation or a cost plan, the pattern is the same.

More questions, and longer ones. A client who once rang with one question can now send ten, numbered and neatly worded, before you’ve finished your first site visit of the day.

Challenges to quotes and instructions. Clients paste your fee quote or your terms into AI and ask whether they’re fair. AI tends to agree with whoever’s asking, so a worried client hears back what they were already worried about. Then you’re asked to justify your quote line by line.

Complaints that take seconds to write. The Property Ombudsman saw a 50% jump in tenant complaints in the first half of 2026, and its Chief Ombudsman, Lesley Horton, says AI is part of the reason [5]. The Financial Ombudsman Service found that as many as a third of replies to its initial assessments, in a small sample, were written or heavily shaped by AI. In its worst example, 37 of the 39 references in one submission were wrong [6].

Claims built on things that don’t exist. In February, a surveyor shared how a rival firm had put their report into ChatGPT to critique it. The AI cited legislation that doesn’t exist [1]. Invented or not, every point in a long AI letter still has to be read, checked and answered.

Writing a complaint used to take an evening and a bit of nerve. Now it takes one prompt.

Defending your work still costs exactly what it always did.

Why the gap is exposure

The rules bind you, not them. The Property Ombudsman has said it won’t make complainants declare whether they used AI, because its codes set standards for agents, not consumers. Instead, it expects agents to keep clear, accurate records [7]. The RICS AI standard works the same way [4]. It applies to surveyors and firms, not to their clients.

Your client’s AI answers to no one. Yours has to be evidenced.

The standard hands clients the questions. Clients can ask how AI was used on their job and challenge the outcome, and firms must be able to explain it in writing, including how they decided the AI’s output could be relied on. A client’s AI can draft that request in seconds. Could your firm answer it by the end of the week?

Your terms are the first place they’ll look. If AI plays a material part in your service, the standard expects your terms of engagement to say so in advance. If your terms are silent, an AI-drafted complaint can point that out in one line.

“We don’t use AI” may not hold. AI built into your reporting software counts if it shapes the work. And RICS has said it will take its standards into account in disciplinary proceedings, as a judge or adjudicator is likely to [4].

What to do about it

1. Speak to clients where you can. An AI-written email invites an AI-written reply, and every written reply becomes part of the record. A phone call gets to the real issue faster, because the real issue is often buried under the AI’s extra detail. The Financial Ombudsman has started picking up the phone for exactly this reason [6]. Afterwards, confirm in writing what you agreed. Ian Quayle covers what that note should say (more below). If a client will only communicate through AI, like the one I opened with, a call isn’t an option. Then your written replies are the record, and each one needs to be right first time.

2. Get your explanation in first. Before a client asks AI whether your quote is fair, tell them yourself: what’s included, what isn’t, and why. Put it in plain English, in the quote itself. AI fills gaps with what’s typical, so leave it fewer gaps.

3. Use AI to handle AI, properly. On the right tool, AI is good at this work. It can cut a long AI-written challenge down to the points it actually makes, list every law and case it cites so someone can check they exist, and draft a first reply for a surveyor to rewrite. Greg Tsuman, a former Propertymark president, says his agency now uses AI to summarise tenant complaints, while every reply is still handled by a person [5].

But a complaint is full of personal data, and the standard is strict about uploading it. Take names and addresses out first, and use a business-grade tool, never a free one.

4. Check your terms of engagement this month. If they don’t cover AI, get them reviewed properly. A clause stopping clients uploading your reports to public AI tools needs legal advice, especially with consumer clients.

5. Make sure you can show your working. When the complaint comes, nobody will ask whether you meant to use AI responsibly. They’ll ask you to show it: what you used, why, who checked it and when.

Your clients aren’t going to slow down. The gap only closes from your side.

Two ways to close the gap