AI can prepare the work. A professional must stand behind it
What the RICS Responsible Use of AI standard means for cost work: faster preparation, structured evidence and a named professional decision.

AI can measure, organise and draft cost work in minutes. Speed is valuable, but it is not the same as reliability. When an output affects a construction budget, tender or client recommendation, someone still needs to understand its inputs, test its assumptions and decide whether it is safe to use.
RICS’ 2026 construction survey shows why this matters now: AI use is spreading faster than organisational readiness. Around two-thirds of construction respondents use AI somewhere, while nearly two-thirds describe their organisation as not prepared or only minimally prepared; fewer than 1% report full integration. The hard work is turning a useful pilot into a controlled, repeatable practice.
The RICS Responsible Use of AI standard, effective from 9 March 2026, formalises that distinction for RICS members and regulated firms. When AI has a material impact on a surveying service, the work needs governance, a professional reliability decision and clear communication with the client. Software can carry much of the evidence; it cannot take over professional responsibility.
Decide where AI is material
Not every AI task carries the same risk, so governance should begin with the effect of the work.
Using AI to organise meeting notes is different from using it to prepare a cost plan, recommend a tender adjustment or assess a payment. The second group can materially affect a surveying service and therefore needs stronger control.
RICS requires regulated firms to record AI systems that could materially affect their services and to assess in writing whether AI is appropriate for the task. Members and firms still need a proportionate, recorded materiality decision. Making that decision before the output is relied on gives the team a route: routine assistance can stay light, while material cost work receives an explicit review.
Do this
- Record the AI system, task and intended use.
- Decide whether the output could materially affect the professional service.
- Document whether AI is appropriate and what additional control the task needs.
Before you move on
Could the firm explain why this use of AI received this level of review?
Give AI structured project evidence
A polished answer is only useful when the project evidence behind it can be inspected.
A structured BOQ gives an AI addressable work items, quantities, units, resources and assumptions. Connected drawings, revisions, rates and project decisions provide context that a copied spreadsheet or isolated chat usually loses.
Keep source facts, calculations, assumptions, suggestions and unresolved gaps separate. Give the AI only the project information and authority needed for the task. Before confidential project information is uploaded to a third-party AI service, RICS expects a risk assessment and the client’s express written consent in advance. Better structure improves the first pass and makes it much easier for a professional to find what still needs judgement.
Do this
- Keep source documents and revisions connected to the affected work.
- Separate facts, calculations, assumptions and unresolved gaps.
- Limit each AI tool or agent to the data and actions needed for its task.
Before you move on
Can the reviewer see what the AI used, what it inferred and what it could not know?
Make reliability a named decision
Material AI output does not become dependable professional work merely because it looks complete.
For an AI output with a material impact, RICS requires a written reliability decision prepared by, or under the supervision of, an appropriately qualified and named surveyor. It should address assumptions, concerns about the input or output, the reasons and mitigations, the effect on the service and the conclusion. High-volume work still remains the firm’s responsibility; RICS calls for randomised, regular dip samples rather than an assumed safe sample size.
The review should focus on the things that can move the decision: scope completeness, quantity sources, rate basis, project conditions, unusual values and unresolved risk. Record who reviewed the work, what they checked, what changed and why the final position was accepted or held.
Do this
- Name the qualified person responsible for the reliability decision.
- Prioritise high-value, unusual and low-confidence outputs for review.
- Record the checks, corrections, decision and approval date.
Before you move on
Could the named professional explain why this output was considered reliable enough to use?
Keep the client and audit trail in view
Responsible AI use must remain explainable after the advice has been issued.
RICS requires relevant client terms to explain in advance when, why and where AI will be used, as well as any available process to contest the use or opt out. Regulated firms also need wider controls: a system register, risk register, supplier due diligence, assigned responsibilities and annual training. The project record should support that process by preserving the system used, its sources, the reviewed output and the accepted professional decision.
A platform can remove much of the practical burden by keeping that evidence with the work and producing a clear review trail. The member or regulated firm still owns its policies, client terms and professional judgement, including explaining the role of AI if asked. That boundary is important: good software helps a firm operate and demonstrate its controls; it does not issue blanket compliance.
Do this
- Explain material AI use in the documents governing the client relationship.
- Keep the AI system, project evidence, review and final decision together.
- Turn recurring corrections into a firm-level risk control or review check.
Before you move on
Could the client and the firm reconstruct how AI contributed without relying on memory?
In short
What to carry into the next job
- AI speed does not remove the need to decide whether a material output is reliable.
- Structured inputs, visible assumptions and controlled access make professional review stronger.
- A named professional retains judgement and accountability for accepted material outputs.
- The platform can carry the evidence and review trail, but the member or firm remains responsible for compliance.
Sources and standards
Read the official guidance
- RICS: AI in commercial property and construction report 2026 ↗
- RICS: Responsible use of artificial intelligence in surveying practice ↗
- RICS: Key elements of the professional standard ↗
- RICS: Client information note ↗
- RICS: Responsible use of AI case study — Construction ↗
- RICS: Cost prediction professional standard ↗
Read next · Guide 01