Context Changes Everything
A photo can tell you a crack exists. Computer vision can tell you what it means.
That's the idea at the heart of our new eBook, AI in Building Surveying: Why Context Changes Everything, out now. It's written for building surveyors and practice principals who are trying to work out what to actually do about AI, not the breathless enthusiasm, not the paralysing fear, but the practical, honest version. Here's a taste of what's inside.
The landscape is shifting faster than most practices realise
The RICS AI Standard came into force on 9 March 2026 and is now a mandatory conduct standard for RICS members. Its central concept is materiality: not every AI tool triggers its requirements, but any AI that materially affects professional output, drafting a valuation, flagging what to investigate, shaping advice to a client, does. Once that threshold is crossed, firms must be able to explain in writing what type of AI was used, its limitations, and how risks are managed.
At the same time, the 2025 RICS Skills Report found that 57% of respondents said skills shortages were already reducing their capacity to take on work. This isn't a problem five years away. It's shaping decisions practices are making right now, including which AI tools are worth their attention.
What surveyors are actually losing time to
The eBook makes a case that's easy to miss: the bottleneck usually isn't the survey itself. A thorough inspection might take three or four hours, but the real time cost comes afterwards, sorting photographs, constructing report language, checking compliance, formatting, sending. That work falls unevenly too. Senior surveyors write fluently because the language is second nature to them; junior surveyors are often doing the same quality of observation but need far longer to turn it into a report that reads consistently.
Structured capture tools already on the market help with this, but only up to a point. They organise the data. They don't understand it.
Why computer vision is a different problem to image analysis
This is the chapter I'd point most people to first. Image analysis, in the sense most AI surveying tools use it, is pattern matching: a model trained to recognise what a crack looks like. It's genuinely good at that. But it doesn't know the age of the building, the construction type, or what else is going on around the defect it's just flagged.
Computer vision, properly built, does. The same 2mm diagonal crack from a window opening means something different in a Victorian terrace than in a 1990s cavity wall, and an experienced surveyor reads that instinctively. The eBook sets out how a system can be built to reason the same way, classifying defects rather than simply matching them to a training image, and why that distinction matters enormously the moment a client questions a recommendation or a dispute arises.
A practical evaluation framework
Rather than a list of features to be impressed by, the eBook sets out the questions worth asking any AI vendor: What was it actually trained on? Is the output auditable? What does it not do? What happens to your survey data? It's built to be used, whether you're trialling a tool now or building a case for one within your practice.
Read the full eBook
This is only a fraction of what's covered. The full eBook goes deeper into the RICS human-in-the-loop framework, what realistic AI adoption looks like for a small practice of surveyors, and a full evaluation checklist you can apply to any tool on the market.
Please enter your details to access the eBook: