Yes, if AI writes only from a checked fact sheet for each unit and its advertising permit record, never from an agent's memory. It drafts the Arabic and English descriptions together, checks every claim against the fact sheet, flags anything it cannot support, and a named person in the listings team approves before the ad goes to any portal.
#Key takeaways
- The risk in listing copy is not grammar; it is a claim the unit or the permit does not support.
- One structured fact sheet per unit is the source for every description, in every language and on every portal.
- AI can draft at volume and check its own draft against the fact sheet, but a person approves what goes live.
- A template and a ChatGPT team plan are enough for a handful of listings; a workflow pays off at portfolio volume.
- Measure drafts approved without edits, listings rejected or corrected, and regulator warnings.
#Why is listing copy a compliance task, not only a marketing one?
In Dubai, property ads need a permit through the Trakheesi system, and the ad has to match what the permit covers. The regulator inspects. Dubai Land Department, july 2024 reported more than 1,200 legal warnings and fines for 256 brokers over a half-year for not complying with advertising terms and conditions.
Enforcement is not new. Dubai Media Office, october 2020 announced fines on ten real estate companies and warnings to thirty more for advertising without valid permits or misusing permit numbers. A description that adds a sea view, a bigger plot or a handover promise is a risk, however good it reads.
#What does a listings coordinator's week look like?
New instructions arrive from agents by WhatsApp, with photos, a voice note and half the details. The coordinator chases the size, the floor and the permit number, writes the English description, sends it for Arabic translation, then copies both into Property Finder, Bayut, the website and the CRM, each with its own fields and character limits.
At volume, the same unit ends up with three slightly different descriptions, an expired permit number on one portal, and an Arabic version that says something the English one does not.
#A listing workflow your team can copy
- Build one fact sheet per unit: reference, permit number and status, type, size, floor, bedrooms, view, furnishing, availability. The agent fills it; the coordinator checks it.
- Write a short style guide in Arabic and English: tone, words you never use, how you describe views and amenities, and what each portal requires.
- Let AI draft both language versions from the fact sheet and the photos' captions, not from free text.
- Have AI check its own draft: every claim traced to a fact sheet field, permit number present and current, nothing promised about future works or handover.
- A named person approves each listing before it is published, and edits go back into the style guide.
- When a unit changes or the permit expires, the fact sheet changes and every live description is flagged for update.
#What should AI never write in a listing?
- A feature that is not on the fact sheet, such as a view, a parking space or an upgrade the agent mentioned in passing.
- Superlatives you cannot prove: best, only, or promised returns and rental yields.
- Anything about handover, future community facilities or payment plans that the developer or the permit does not state.
- A permit number it was not given, or one that has expired.
#When is a template or a general AI tool enough?
If you list a few units each week, a good template, a careful agent and a general assistant such as ChatGPT or Copilot are enough. Paste the fact sheet, ask for both languages, check it line by line. Many portals also offer description helpers that cover the basics.
A workflow earns its place when you publish at portfolio or project volume, run several portals, and need one record of who approved which description. For multi-channel content more broadly, see AI for content production.
#Where does Stennir fit?
Stennir usually meets listings inside a wider review. AI consulting ranks five workflows across the brokerage, with a cost and payback case for each, and recommends one build. If listings rank first, AI implementation builds the fact sheet, the bilingual drafting and the claim check into the CRM or listings tool you already use, with your coordinator approving every ad.
AI training then works with agents and coordinators on the real listings, so fact sheets arrive complete and drafts need fewer edits. Every action is logged with who approved it, which the NIST AI risk management framework, january 2023 treats as part of managing AI risk.
We are
Stennir builds a listing workflow where every description traces back to a checked fact sheet and a named person approves before anything is posted.
We aren't
Stennir is not a portal, an advertising permit service or a legal adviser on what your regulator allows.
#What should a head of listings do next?
Pull ten live listings and check each claim against the unit record and the permit. The gaps you find are your case. See AI for sales teams, then take the free readiness assessment or book a 30-minute call.
#Common questions
Is the Arabic written in Arabic or translated from English?
Both versions are drafted from the same fact sheet, so the Arabic is written for Arabic readers rather than translated line by line. Your style guide sets the tone in each language. A person who reads Arabic approves the Arabic version, and differences between the two versions are flagged before either is posted.
Does AI check the advertising permit?
It checks that a permit number is present on the fact sheet, that the record you hold shows it as current, and that the ad stays within what that record covers. It does not replace your own check with the regulator. A named person still confirms the permit before the listing goes live.
What do you need from us?
Your listing fields, a sample of past descriptions you are proud of and some you are not, the portal requirements you follow, and one person who owns the style guide and approves listings. Access to the CRM or listings tool is agreed with your team before anything is connected.
What could go wrong?
A fact sheet with a wrong field produces a confident, wrong description, and a style guide that allows superlatives produces ads that overclaim. That is why the fact sheet is checked by a person, the claim check flags anything unsupported, approval is required before posting, and every change is logged.
