Skip to content
secruna
Get a Demo
RICS standard live since September 2025
RICS Responsible Use of AI in Surveying Practice

Show your RICS regulator you’ve got AI under control.

The RICS Responsible Use of AI in Surveying Practice standard (September 2025) requires every regulated firm to keep an up-to-date register of AI tools, disclose AI use to clients per instruction, and retain an audit trail a Standards & Regulation review can read. Secruna discovers the tools across your stack, classifies them under the standard, generates the disclosure statement, and keeps the evidence — so your next firm review is paperwork you already own.

Get a 30-min RICS scope call
Why this matters

Three exposures bite — in this order.

The RICS standard is already in force. The pressure on a surveying firm comes from three directions at once — and each of them gets harder to answer the longer the AI register sits unstarted.

Regulatory exposure

The RICS Standards & Regulation team can open a conduct review under the Rules of Conduct and the Responsible Use of AI in Surveying Practice standard (PS, September 2025, § 3 firm-level obligations). A finding of non-compliance carries CPD remediation, public censure and, at the upper end, suspension of the firm’s regulated-services status. The standard is binding on every MRICS / FRICS practitioner and every regulated firm — it is not advisory.

Client trust

Clients of regulated services — institutional landlords, lenders, conveyancing solicitors, pension-fund property arms — increasingly require a disclosure of AI use upfront with the instruction. Without a register and a disclosure statement, you cannot answer the question on the proposal, and you lose the instruction before the fee is even quoted. The first firms to ship the disclosure win the trust premium for the next 24 months.

Professional indemnity

Surveying PII insurers are rewriting renewal questionnaires to ask which AI tools are in use, how the firm supervises their output, and whether disclosure to the client was made. Without contemporaneous evidence the renewal premium goes up — or, in a hardening market, cover is refused. The renewal cycle catches every firm in 12 months. A clean register at renewal is the cheapest underwriting argument you can make.

The five-step path

What you have to do, in order.

The RICS standard imposes the same five gates on every regulated firm. Start at step one — the rest only make sense once you know which AI tools your surveyors actually touch.

  1. 1

    Inventory

    Discover every AI tool in your tech stack — CoStar AI, Argus Enterprise, Reapit, Yardi, Aprao, VTS, plus the ChatGPT and Copilot integrations sitting inside Microsoft Word and Excel. Most firms underestimate the count by 3× on the first scan; the second scan, run a fortnight later, almost always surfaces another tool a single team enabled without telling compliance.

  2. 2

    Classify

    Map each tool to the relevant RICS PS Sep 2025 section — disclosure required, supervision required, professional judgement only, or out of scope. A single tool can sit in more than one bucket depending on the instruction it is used on; the classification follows the instruction, not the licence.

  3. 3

    Disclose

    Generate the AI Use Disclosure Statement per instruction — the exact wording the chartered surveyor signs alongside the report. Secruna pre-fills the statement from the firm register so the surveyor edits a draft, not a blank page, and the client sees a consistent disclosure across every report your firm issues.

  4. 4

    Register

    Maintain the Firm AI Register that a RICS firm-level review will ask for: tool name, vendor, RICS category, named senior responsible owner, last-reviewed date, client-disclosure status, retention setting. Secruna keeps the register live as new tools enter the stack and flags entries that have aged past their review date.

  5. 5

    Audit trail

    Retain six years of evidence in line with the RICS retention norms — every classification change, every disclosure version, every reviewer sign-off. When the Standards & Regulation team opens a review, the firm-level pack is one click away rather than three weeks of email archaeology.

Automated valuation models (AVMs)

Where does an AVM end and chartered judgement begin?

The buyer’s question. A residential valuer uses CoStar AI or a lender’s in-house AVM to produce a comparable-sales valuation. The output materially shapes the surveyor’s figure. What does RICS require the firm to do, and what does the surveyor sign?

RICS PS Sep 2025 reference. § 4.1 on automated valuation models classifies AVM output as an aid to chartered judgement, not a substitute for it. The chartered surveyor remains personally responsible for the figure that goes into the report; the AVM run, its inputs, its comparable-evidence set and its result must be retained as part of the instruction file.

What counts as compliant. Disclose AVM use in the report. Document the comparable-evidence set the model used (not just the headline figure). Sign-off by an MRICS surveyor. Retain inputs and outputs alongside the instruction file for the standard six-year retention window. An AVM whose output is signed without a comparables record fails the standard regardless of accuracy.

What Secruna ships for AVMs. Auto-detection of CoStar AI, Reapit valuation modules, Yardi-Voyager valuation extensions, and the open-market AVM APIs your team integrates with. A pre-filled AVM disclosure block in the AI Use Disclosure Statement. A retention rule tied to the instruction so the comparables snapshot does not age out before the six-year window closes.

See this in your dashboard at: /rics/tools?category=requires_supervision filtered to AVM tools, with the disclosure-statement preview on each card.

AI-assisted GIS & mapping

When a flood-risk overlay uses AI, does the client need to know?

The buyer’s question. A building surveyor uses an AI-driven flood-risk layer or an automatically classified land-use overlay to inform a condition report or a development-feasibility opinion. The GIS layer was generated by a model, not a human. Does the firm have to tell the client?

RICS PS Sep 2025 reference. § 4.2 on AI-assisted geospatial tooling treats AI-generated boundaries, flood-risk classifications, contamination-risk tiles and similar geospatial layers as inputs to the surveyor’s professional opinion. Where the AI layer materially affects the conclusion, the client must be informed — that is, disclosure is required, not optional.

What counts as compliant. Identify the GIS layers in use, name the ones that are AI-generated, record which instructions used them, and add a one-sentence disclosure to the report explaining the AI source of the layer. The disclosure is not buried in a footnote — it sits with the rest of the assumptions and caveats.

What Secruna ships for GIS. Detectors for the common GIS platforms with AI overlays (Esri ArcGIS Online, Landmark Promap, MapInfo, Future Climate Info), an AI-layer flag on each tool record, and a reusable disclosure paragraph the surveyor pastes into the Assumptions section of the report.

See this in your dashboard at: /rics/tools?category=requires_disclosure filtered to the GIS tools your firm has connected, with the disclosure paragraph available for one-click copy.

Generative AI in report drafting

ChatGPT in Word — helpful tool or supervised output?

The buyer’s question. A surveyor uses ChatGPT, Microsoft Copilot, Claude or Google Gemini to draft an executive summary, a condition narrative, or the introductory sections of a report. The chartered surveyor edits the draft, signs the final version, and issues it under the firm’s name. Does that count as an AI use the standard governs?

RICS PS Sep 2025 reference. § 5.1 on generative AI in client-facing surveying output treats any AI-drafted text the surveyor adopts into the report as a supervised use — the surveyor remains personally responsible for the words that go out under their name. Identical fact pattern often also triggers EU AI Act limited-risk transparency for clients in EEA jurisdictions; both frameworks fire on the same use.

What counts as compliant. The chartered surveyor reviews and adopts each AI-drafted section before it ships. The firm records that generative AI was used in drafting (not necessarily which prompt — the standard does not require prompt logs). The disclosure statement makes clear AI was used for drafting and that the surveyor accepts professional responsibility for the final text.

What Secruna ships for drafting. Detection of Copilot for Microsoft 365 seats, ChatGPT browser extensions on managed devices, Claude API integrations into Word add-ins, and Gemini for Workspace. A pre-built drafting-disclosure paragraph in the AI Use Disclosure Statement that the surveyor adapts per instruction.

See this in your dashboard at: /rics/tools/generative-ai for the consolidated view of every drafting-AI tool the firm uses, who has access, and the live disclosure language.

Document extraction in due diligence

When a model reads the lease, who is responsible for the answer?

The buyer’s question. A commercial surveyor runs an AI extractor across leases, title deeds, planning permissions, EPCs and condition reports to lift key facts — break dates, rent reviews, service-charge caps, restrictive covenants — into the surveying decision. Extraction errors are silent. What does RICS require the firm to do?

RICS PS Sep 2025 reference. § 5.3 on AI-assisted document extraction in due diligence treats every extracted fact that influences a surveying decision as supervised output: a chartered surveyor must verify the extracted fact against the source document before it lands in the report. Acceptable sampling rates are explicitly addressed — 100% verification for material facts, defined sampling for the rest.

What counts as compliant. Identify which extraction tools are in use (Leverton, Imprima, Kira, Eigen, Luminance, in-house Claude pipelines). Define a firm-wide policy on which fields are material and require 100% verification. Record the verification — who checked, when, against which version of the source. Retain the evidence with the instruction file.

What Secruna ships for due diligence. Detection of the major DD extractors plus the in-house Claude / GPT-4 pipelines that show up in your Azure or AWS account. A verification-policy template pre-populated with RICS-aligned material-field definitions, and a per-instruction sign-off log that proves the surveyor actually checked rather than rubber-stamped.

See this in your dashboard at: /rics/tools/extraction for the tool list, and /rics/policies/verification for the firm-wide verification policy you sign off once and apply everywhere.

Client-facing chatbots

Your website chatbot just answered a surveying question. Now what?

The buyer’s question. The firm deploys an AI chatbot on the public website, the client portal, or as a WhatsApp Business assistant. A client asks a question a chartered surveyor would normally answer — about a damp survey, an EPC, a Schedule of Condition. The chatbot replies. Has the firm just given a regulated opinion?

RICS PS Sep 2025 reference. § 5.4 on AI chatbots in client-facing channels treats any reply that a reasonable client could read as a surveying view as falling within the disclosure regime. The firm must disclose that the responder is AI, must route any substantive surveying question to a chartered surveyor, and must keep a transcript long enough to support a complaint or a regulatory query (six years).

What counts as compliant. A clear AI disclosure on the first interaction (not buried in the terms of service). A topic-routing rule that escalates substantive surveying questions to a named human within a defined SLA. A transcript-retention policy that survives the chatbot vendor swap. An opt-out path that lets the client speak to a person without first triaging through the bot.

What Secruna ships for chatbots. Detection of the common bot platforms (Intercom, Drift, HubSpot, custom OpenAI / Anthropic deployments fronting your website), a pre-written first-interaction disclosure block, and a transcript-retention policy that ties to your firm-level retention setting rather than the vendor’s default.

See this in your dashboard at: /rics/tools/chatbots for every client-facing bot, its disclosure status, and the retention setting you have applied.

See where your firm stands.
In 30 minutes.

A 30-minute scope call gives you a concrete answer for each of the five RICS rules above — which of the AI tools your surveyors use today fall under which obligation, and what evidence is missing before the next firm-level review.

Get a 30-min RICS scope callSee the EU AI Act page

Or call our UK lead — we’re on +44 20 0000 0000. (Placeholder — see TODO at the top of this file; the real number lands once the founder confirms it.)