How SurveyorSuite handles AI — a surveyor's guide

What AI does in SurveyorSuite

Only one feature uses generative AI — the one that matters for RICS §4.2 compliance. Everything else — photos, measurements, layout, calculations — is deterministic code.

Generative AI (in scope for RICS §4.2)

• Field Notes → report drafting. Paste your on-site notes when creating a new survey. Anthropic's Claude converts them into structured report content, populates the relevant fields, and applies your chosen rating codes. You review, edit, and sign off before export.
• Writing-style tuning. Phrases you mark as favourites in your phrase library are sent as tone examples alongside your notes, so the generated draft reads in your voice rather than a generic "AI voice". This sits inside the Field Notes feature — no separate setting.

Speech-to-text transcription (not generative AI)

SurveyorSuite also offers dictation on every text field. This is a speech-to-text transcription — your own words, typed out automatically. It does not generate new content, reason about your input, or take professional decisions, so it is not covered by the RICS §4.2 reliability-decision requirement. It is covered by §3.1 data-protection duty and §4.3 Terms of Engagement disclosure, both of which we support.

• On the web app — uses your browser's built-in Web Speech API (Google for Chrome, Microsoft for Edge, Apple for Safari). Audio goes directly from your browser to your browser vendor; we never see it. Only the transcribed text is saved.
• On the mobile apps — uses your device's built-in speech recognition (Apple SFSpeechRecognizer on iOS, Android RecognitionService on Android). Runs on-device where the platform supports it; audio is not stored by SurveyorSuite.

What AI doesn't do

The surveyor is always the accountable professional. The AI does not:

• Auto-export reports. Every PDF and DOCX is issued manually by the named surveyor.
• Make final professional decisions. It drafts — the surveyor reviews, edits, and signs off.
• Run without disclosure. If AI was used, the exported report includes an AI Disclosure block naming which categories of output (text, photo analysis, research) were AI-assisted and a reliability attestation from the surveyor.
• Train Anthropic's Claude on your data. Your field notes and reports are never used to train Anthropic's models — that is an explicit commitment in their Commercial Terms. Speech-to-text uses your browser or device's speech APIs (Google / Microsoft / Apple); their handling of voice input is governed by their own privacy policies, so check those if voice data is sensitive for you.

What surveyor users need to do for RICS compliance

The RICS Professional Standard Responsible use of artificial intelligence in surveying practice (effective 9 March 2026) places the responsibility on you, the regulated member or firm — not on the software. SurveyorSuite is designed to make that easier. You still need to:

• Apply professional judgement before accepting any AI output. Every AI-assisted section should be reviewed on its own merits — not rubber-stamped.
• Include AI disclosure in your Terms of Engagement. We provide default RICS-aligned clauses (covering AI use, scope, PI cover, opt-out, contest and redress) in two places depending on how you engage clients: in Settings → Report Terms if you send ToE by email yourself, and in the CRM's Terms Templates if you're on a Practice plan where clients sign via the client portal. Both surfaces draw from the same canonical clause set and can be adapted to your firm's language.
• Don't paste confidential client context into AI. Health information, financial detail, disputes, divorce, or any personal data the client hasn't given you written consent to share with an AI system — keep it out of the Field Notes box.
• Keep a firm-level AI register and risk register. An AI system + risk register tool is coming to Settings for firm admins — meanwhile, contact support if you need help with your RICS compliance file.

What SurveyorSuite does behind the scenes

• Model. AI requests go to Anthropic's Claude family of models via their commercial API. We may vary the model version per template to match task complexity; this does not change what is disclosed in your exported report.
• No training on your data. Anthropic does not train their models on data submitted via the API. Their Commercial Terms of Service state (§B, Customer Content, effective 17 June 2025): "Anthropic may not train models on Customer Content from Services." This applies uniformly across direct API, Amazon Bedrock, and Google Cloud Vertex AI.
• Audit trail. Every AI generation is logged in an audit trail tied to the surveyor's account. The log is visible to the surveyor and to SurveyorSuite platform staff, and is retained alongside the report record for PI-insurance retention periods.
• Surveyor sign-off. A qualified surveyor reviews and signs off every AI-assisted section before the report is exported. The exported report's AI Disclosure block includes a reliability attestation to this effect.
• Transient handling at Anthropic. Prompts and responses are held transiently by Anthropic for abuse monitoring per their API terms; we do not retain them beyond the audit trail shown above.

Supplier Due Diligence Pack

If your firm's compliance officer needs documentation for their §4.1 procurement file, we provide a one-page Supplier Due Diligence Pack covering environmental impact, training data provenance, liability terms, and data-protection confirmations.

Download the Supplier Due Diligence Pack (PDF) →

If the download link is not yet live, contact [email protected] and we will send the current version directly.

Sources and verification

Every factual claim on this page is verifiable against primary sources. If your firm or client needs to confirm anything we say before signing Terms of Engagement, use the links below.

• Anthropic Commercial Terms of Service — anthropic.com/legal/commercial-terms . The no-training-on-your-data commitment is in §B (Customer Content), effective 17 June 2025.
• Anthropic Privacy Center — commercial customers — privacy.claude.com . Summary of data-handling practices that apply to API / Claude Enterprise customers (different from consumer Claude.ai).
• RICS Professional Standard — Responsible use of artificial intelligence in surveying practice, 1st edition (Sept 2025), effective 9 March 2026. Free to download from rics.org (search: "responsible use of artificial intelligence"). ISBN 978 1 78321 555 3.
• In-app help → AI & Voice-to-Text section — accessible from any SurveyorSuite survey editor (surveyor-facing reference).

If a link above is broken or the source has moved, email [email protected] and we will update this page.