A practical guide / Estate and letting agents
What AI do estate agents use?
Most estate and letting agents use AI in two places: features inside their property CRM and business AI assistants such as ChatGPT, Copilot or Claude. Estate and letting agents can use AI to draft listings from checked property facts. It can also summarise viewing feedback and sort tenancy tasks in a shared tracker. Start with the fact sheet, then choose the output you need.
Give the listing two parts. Write a clear description, then add the factual schedule. Keep both linked to the same property ID. This gives the agent a clear basis for the advert, viewing conversation and follow-up.
Draft property particularsWrite particulars and updates from the same property record.
Give every listing a property fact sheet
Build the fact sheet before writing the description. Include the address, asking price or rent, rooms, measured sizes, outdoor space and logged parking details. Add tenure and charge details where relevant. Link the source documents. Give each fact a source. Add the date checked.
More on this step
Ask AI to draft from this sheet. Write an opening paragraph, room descriptions and the factual schedule. Set the tone to match your agency and choose a length for the portal. A useful instruction is to keep qualifiers such as “allocated” or “shared” when they describe a feature.
Keep a separate question list for missing information. When the agent resolves a question, update the fact sheet and regenerate the affected passage. Update both the portal text and printed sales details from this sheet. Keep them aligned as facts are confirmed.
From property facts to readable sales details
Example with made-up details: the record describes a two-bedroom first-floor flat. It has a separate kitchen, shared rear garden and one allocated parking space. The station distance and lease figures are awaiting reply.
More on this step
The opening reads: “This first-floor flat has two bedrooms and a separate kitchen. It has access to a shared rear garden and an allocated parking space.” The rooms section then follows the measured floor plan, with each room described in order.
The preparation list asks the agent to check the lease information and station distance. Once those facts are available, the team can complete the relevant fields. The draft already gives a potential viewer a clear account of the features that have been established.
Turn viewing comments into a useful seller update
Give each viewing a date and ID. Use headings for layout, condition, price and location. Add the next step. Ask AI to summarise repeat themes. Show how many viewers mentioned each one. Keep a single person’s comment attributed as one response.
More on this step
Use three headings for the seller update: response so far, themes from viewings and decisions to discuss. A theme such as kitchen condition can sit beside the relevant feedback extracts. The agent can then add their own advice. Cover presentation, follow-up or price as needed.
Close with an action list: viewer to contact, question to answer, owner and agreed date. This turns the weekly update into the next set of tasks. A colleague covering the property can follow the calls already under way.
A tenancy tracker that shows the next action
Create one row per tenancy task, using property ID, task, required document, owner, target date and current status. Set the required tasks and dates from your agency’s current process for the relevant nation. Use AI to sort the records and prepare a list of items awaiting action.
Keep document expiry dates separate from visit dates and internal chase dates. These represent different events. A booked visit can have its own entry while the resulting certificate remains pending.
Use the tracker to prepare a daily work list for the lettings team. Group tasks by the person responsible and show what they are waiting for. Add a link to the fact sheet so staff can open the relevant document or messages during the call.
Publish accurate information under current guidance
GOV.UK’s consultation on material information covers residential sales. It explains the duty to give consumers the facts they need. It also proposes new guidance. It is a consultation, so review current official guidance when setting your agency’s listing checklist. Apply the rules for the nation and transaction involved.
GOV.UK explains when an Energy Performance Certificate is needed for a sale. Add the applicable EPC details to your listing check. The ICO’s AI guidance covers personal data accuracy and security. Use approved tools for viewing notes and tenancy records. Have the responsible agent approve the final listing against the property documents.
Read the source: GOV.UK: material information consultation for property listings; GOV.UK: selling a home and EPCs; ICO: guidance on AI and data protection
Try a practical task
Draft sales details from our checked fact sheet
Write sales details for [property ID] using the attached fact sheet and measured floor plan. Draft an opening of up to 70 words. Describe rooms in floor-plan order, then fill in our portal fields.
Try this in your AI app
ChatGPT
Claude
Gemini
CodexUsing Gemini? The button copies the prompt. Paste it into the new chat.
Keep the exact descriptions of tenure, shared areas, parking and charges. Use only measurements and distances present in the record. List unanswered fields for the agent. Name the document or contact needed for each one.
Give each factual passage a source ID in the working draft. Finish with a short version for a viewing confirmation email, using the same checked features. Maintain our agency’s plain, factual tone throughout.
Questions about this task
Can AI write the whole property advert?
It can prepare the description and factual schedule from a complete fact sheet. Give the agent the source-linked working draft and a list of outstanding fields to resolve before listing.
What is a useful first task for a small agency?
Try a viewing feedback summary for one property. Use the dated comments to find themes, prepare the seller update and assign follow-up calls. This gives the team a useful output from information it already records.