Skip to content
Curious Context Labs

case study · Marketing · Custom app

Accurate, distinct property listings the morning after the viewing

How an agent's voice note, floor plan and photos can become portal-ready copy, with every feature checked before it goes live.

business
Real estate
build
Custom app
timeline
3 weeks
read
3 min

01 the business

Independent estate agency

4 branches, 22 agents

Agents write listing copy late at night from memory and notes. Descriptions sound alike, and details such as parking or room sizes are sometimes wrong. Wrong details create complaints and, in some markets, legal risk.

02 before

The problem

Agents write listing copy late at night after viewings, from memory and scribbled notes. Descriptions read the same across properties, and details such as parking or room sizes are sometimes wrong. Holiday lets and hotels face the same job for every room type.

  • Listings take two or three days to go live after instruction.
  • Descriptions repeat the same phrases across properties.
  • Errors in sizes or features are found by buyers at viewings.

03 the build

What the AI does

  1. Step 1: Takes the agent's voice note or bullet points, the floor plan and the photo set for a property.
  2. Step 2: Pulls room sizes, features and council tax or strata details from the floor plan and CRM record.
  3. Step 3: Drafts a headline and description for each portal, within its length limits and in your agency's tone.
  4. Step 4: Checks every stated feature against the notes and photos and flags anything it cannot support.
  5. Step 5: Saves the approved copy back to the CRM, ready to publish to the portals.

paused: waiting for a person to approve

How it was put together

  • Took the agent's voice note, the floor plan and the photo set for each property.
  • Pulled room sizes and features from the floor plan and the CRM record.
  • Drafted headlines and descriptions within each portal's length limits.
  • Checked every stated feature against the notes and photos and flagged anything unsupported.

04 human in the loop

Where a person stays in charge

The listing agent approves the final copy before it goes to any portal. Flagged claims must be confirmed or removed first.

  • The listing agent approves the copy before it goes to any portal.
  • Flagged features must be confirmed or removed.
  • The tool avoids descriptions of the neighbourhood that could breach fair housing or advertising rules.

05 rollout

From kickoff to live in 3 weeks

  1. Week 1Gather style examples and portal rules; agree the review screen.
  2. Week 2Draft the next 20 listings alongside the agents.
  3. Week 3Use for every new instruction across branches.

06 after

What changes

Listings go live the day after instruction with accurate, distinct copy. Agents spend their evenings on calls rather than on writing.

worked example · your numbers will differ

Assume 60 new listings a month and 45 minutes to write each one, against 10 minutes to review a draft.

60 × 45 minutes = 45 hours a month of evening writing. 60 × 10 minutes = 10 hours of review. 35 hours a month go back to the agents.

What to measure

  • Days from instruction to live listing.
  • Corrections requested after publishing.
  • Agent time per listing.
  • Enquiries per listing in the first week.

07 evidence

What the research says

  1. In a study of 453 professionals doing writing tasks from their own jobs, access to ChatGPT cut the time taken by 40% and raised quality, as rated by evaluators, by 18%.

    Science / MIT, Experimental evidence on the productivity effects of generative AI (Noy and Zhang) (opens in a new tab) · 2023

  2. The FTC's 2024 rule bans fake reviews, including AI-generated ones, undisclosed reviews by company insiders and threats used to suppress negative reviews.

    US Federal Trade Commission, Final rule banning fake reviews and testimonials (opens in a new tab) · 2024

08 lessons

Mistakes to avoid

  • Flowery copy that overstates. Accuracy sells and protects you.
  • Relying on photos alone for measurements. Use the floor plan.
  • Skipping the agent's own voice. Their notes are what make each listing distinct.

09 questions

Common questions

Does it work for holiday lets and hotels?
Yes. It drafts copy for each room type or property in the same way.
Which portals does it support?
Any portal that takes text from your CRM, such as Rightmove, Zillow or realestate.com.au.

keep reading

All case studies ->

Tell us how your business runs. We'll tell you where AI fits.

A free 30-minute call. You talk, we ask questions, and you leave with a few concrete ideas — or an honest answer that AI isn't the right fix yet.

Working with teams around the world.