AI-enabled surveying tools are changing practice fast. What happens when they stop working?
During the AI surveying tool failure It is the moment nobody plans for. The valuation software goes down. The AI-assisted quantity take-off tool is unreachable. The cloud platform your team has relied on for three months is returning a 503 error. A client is waiting for a report. What do you do?
Most construction and surveying firms do not have a clear answer to that question. They should.
The dependency is real and growing
Surveyors are now using AI tools to take quantities from 2D and 3D models, run automated valuations, check compliance and generate reports. Law firm DWF notes explicitly that "AI design or analysis tools within Building Information Modelling and geographic information systems used by consultants in the construction industry" now sit within the professional liability risk landscape.
RICS has responded. Its new standard on the responsible use of AI, effective March 2026, sets out requirements for risk management, professional upskilling and clear client communication when AI is in use. The standard does not prohibit reliance on AI tools. But it does make the professional responsible for the output, regardless of what the software says.
AI Tool Risk: What the Numbers Say
94% : of IT leaders concerned about AI vendor lock-in (Parallels 2026 survey)
£7,000 : Cost of downtime per minute for large enterprises (Digital Craftsmen, 2025)
44% : of companies report brand damage following a cloud outage
57% : fear future support issues with single-vendor AI contracts
Vendor lock-in is not hypothetical
A 2026 survey by Parallels found 94% of IT leaders are concerned about vendor lock-in, with nearly half describing themselves as "very concerned." The top drivers: uncertain product roadmaps and fears about future support. For surveying firms running workflows through a single platform, this is not an abstract risk. It is a business continuity problem.
Cloud outages happen more often than providers would like to admit. AWS and Azure have both experienced regional failures in recent years, taking downstream services with them. Downtime for SMEs costs between £6,000 and £19,000 per hour. For a surveying practice mid-report, it also damages professional reputation.
Model version control adds another layer of risk. When an AI vendor updates its model, outputs can change without warning. A valuation generated in October may not be reproducible in the same way in January. If your professional indemnity claim depends on being able to recreate how a figure was reached, this matters.
"It remains essential that every AI-generated output is checked by a human. There is no substitute for critical thinking and professional scepticism." DWF, September 2025
What a contingency plan actually needs
Firms that take this seriously are building AI contingency plans into their practice management. Here is what that needs to cover.
First, know your dependencies. Map every AI tool in use to the workflows that rely on it. Which outputs go directly to clients? Which feed into regulatory submissions? Those are your highest-risk dependencies.
Second, define fallback procedures for each. This does not have to mean returning to spreadsheets. It might mean maintaining a parallel manual process for critical outputs, or ensuring a second provider could deliver the same data. For AI-assisted valuations, it means knowing which elements a surveyor can complete manually within the required timeframe if the tool is unavailable.
Third, address version control. Keep records of which model version was used for each output. If your vendor pushes an update, know when it happened. Some professional bodies are beginning to require this as part of responsible AI documentation.
Fourth, check your professional indemnity cover. Many PI policies were written before AI played a substantial role in professional outputs. Insurers are tightening wordings. A conversation with your broker about how AI use is disclosed is overdue if you have not had it.
The deeper issue
Dependency on AI tools is not inherently a problem. The tools are often better and faster than the manual alternative. The problem is dependency without awareness. Firms that cannot name their AI vendors, cannot describe their fallback process and cannot say what oversight they apply to AI outputs are carrying professional risk they have not priced.
The question is not whether to use AI in surveying practice. It is whether you are in control of it, or it is in control of you.
Building AI Competence in the Built Environment
The Responsible with AI programme gives built environment professionals the tools, frameworks and practical guidance to govern AI confidently. Including business continuity planning, vendor assessment and professional indemnity considerations.
Explore the Programme → Responsible with AI
