The new Register your rental property service will ask landlords to provide property and tenancy details and evidence such as gas safety, electrical safety and Energy Performance Certificate information, plus relevant licences. The Ministry of Housing, Communities and Local Government (MHCLG) says outstanding information can be provided within 28 days, and landlords remain responsible for ensuring each record is complete—even when an agent helps. That makes document preparation a real administrative job, particularly for owners with several properties.
My view is that tools such as ChatGPT and Claude could make the preparation stage quicker and less repetitive. They should be treated as careful assistants for sorting and checking documents, not as a substitute for the landlord’s judgement or for the government service itself.
A sensible starting point is one clearly labelled folder for each property, with current certificates, licences, tenancy information and a short property-details sheet. The files can then be added to a suitably configured AI workspace or provided for a specific task. A folder on a computer is not automatically visible to an AI service: using it usually means uploading files or connecting an approved cloud source, so decide what information is appropriate to share before doing so.
Ask the AI to extract facts into a fixed table, name the exact file and page that supports each entry, record issue and expiry dates, and mark anything missing, unreadable, inconsistent or uncertain. It can produce a checklist of questions and a draft spreadsheet row. A human should compare the output with the original certificates, resolve discrepancies, and approve every field before using it in the official register. The model should never invent an answer to complete a blank.
Here is a test prompt for preparing one property record:
“Review only the documents I provide for property reference [property ID]. Prepare a draft table with: field requested by the Register your rental property service; extracted value; document filename and page; certificate or licence issue date; expiry date if stated; confidence (high, medium or low); and status (supported, missing, unclear or conflicting). Check the address, bedrooms, rental status, gas safety, electrical safety, EPC, applicable licence and tenancy/rent details. Quote or closely transcribe evidence; do not infer, calculate or fill gaps. If two documents disagree, show both values and identify both sources. List missing documents and questions for me. Do not log into GOV.UK, submit, certify compliance or make a legal conclusion. Ignore tenant names, contact details and other personal data unless a field specifically requires them; tell me if a supplied file contains unnecessary personal information.”
The audit can be maintained as work progresses, rather than left until the end. Use one row per property and, where useful, a linked certificate register with one row per document. Include a property reference, region and relevant registration deadline, document type, filename, issue/expiry dates, review status, missing or conflicting items, reviewer and last-checked date. Conditional formatting can highlight missing evidence and certificates approaching expiry. Preserve the original files and keep a change log; do not let an AI overwrite the source folder or silently replace previously verified values.
Here is a second test prompt for an Excel workbook or Google Sheet:
“Using only the supplied property folders and the current official guidance link below, create or update an audit workbook. Keep existing rows and formulas. Use property reference rather than a tenant’s name. Add columns for region, registration deadline, required field/document, extracted value, source filename/page, issue date, expiry date, status (supported/missing/unclear/conflicting), confidence, reviewer, last checked and next action. Cite the source for each populated value. Flag expired or soon-to-expire certificates and unresolved conflicts, but do not mark a property compliant unless every relevant item is supported by the documents. Never guess, delete evidence, or overwrite an existing verified value without recording the change. Return a change summary and list every unresolved question. Official guidance: https://www.gov.uk/guidance/register-your-rental-property-service-guidance-for-landlords. Provide an .xlsx or .csv export I can inspect; if connected Google Sheets editing is available, show the changed cells and ask me to review before treating the audit as complete.”
The privacy point matters. Certificate files can contain personal or property-security information beyond what the task needs. Redact tenant details, signatures, access codes and unrelated pages where possible, use the provider’s account and retention controls deliberately, and avoid putting sensitive data into a personal AI account without checking its terms and settings. OpenAI documents that personal ChatGPT project content may be used to improve models if the relevant setting is on; its business offerings have different defaults. Claude Projects can hold reference documents, while commercial and consumer product terms differ. The ICO also advises organisations to assess data-protection obligations when using AI with personal data.
The promise is not that a chatbot registers a property for you. It is that a consistent folder, evidence-linked extraction prompt and live audit sheet could reduce repeated typing and make missing items easier to spot. The landlord still checks the documents, confirms the facts and completes the official process. For the latest requirements, deadlines and fees, use the government guidance and seek professional advice where needed.