Blog · 4 min read
AI skipped real estate firms on at least 87% of "which" questions
By Simplee Editorial Team. Reviewed by Rafael Rositsan · Published · Updated
Sources checked November 9, 2026 · Methodology
When 59 real estate businesses ran our free AI visibility audit between April and October 2026, AI engines skipped the audited business on at least 87% of the 392 "which [provider] should I choose" questions asked across 58 of them (Simplee Digital first-party data, doiexist.ai).
Key figures
- On at least 87% of 392 which-provider questions asked across 58 audits of real estate businesses, April to October 2026, AI engines skipped the business (Simplee Digital first-party data, doiexist.ai).
- At least 49% of 59 real estate businesses audited from April to October 2026 were never named by ChatGPT, Claude or Gemini (Simplee Digital first-party data, doiexist.ai, n=59).
- In audits of real estate businesses from April to October 2026, the trusted or best-reviewed question failed least: at least 67% of 49 times, across 36 audits (Simplee Digital first-party data, doiexist.ai).
A buyer who asks an engine which provider to choose is deciding who to call. In our audits, most of the real estate businesses were not on the list the engine gave back.
Real estate businesses miss the shortlist question most often
Each audit asks ChatGPT, Claude and Gemini the questions a buyer would ask, written to leave the business's name out, and a question fails when no engine names the audited business. The methodology page sets out the rules, and the AI visibility benchmark carries every industry group.
Questions that contained the business's name or its domain were excluded, which leaves 59 audits (Simplee Digital first-party data, doiexist.ai, April to October 2026). In a hand check of 50 stored answers across 5 audits, 45 matched the automated flag. In the other 5 the flag said the business was named but the stored text did not show it: 1 where the flag counted a generic word in the business's name, and 4 in answers cut off at 2,000 characters, where a later mention cannot be ruled out (October 2026).
| Item | Failure rate, at least (see method) (%) |
|---|---|
| which [provider] to choose (392 asked), at least 87% | 87% |
| best for [customer type] (68 asked), at least 83% | 83% |
| price (115 asked), at least 83% | 83% |
| how to find or choose (287 asked), at least 78% | 78% |
| best in [city or region] (298 asked), at least 74% | 74% |
| trusted or best-reviewed (49 asked), at least 67% | 67% |
Source: Simplee Digital first-party data, doiexist.ai, n=59 · Methodology
The shortlist question failed at least 87% of 392 askings across 58 audits, ahead of the next two: best-for questions at least 83% of 68 across 26 audits, and price questions at least 83% of 115 across 45 audits (April to October 2026). All three are what a buyer asks once the search has narrowed.
At least 49% are never named by any engine
In the same 59 audits from April to October 2026, at least 49% of real estate businesses were never named by any of the three engines in any answer (Simplee Digital first-party data, doiexist.ai, n=59). Each engine on its own left out more: Gemini at least 57%, ChatGPT at least 55% and Claude at least 55%.
E-commerce brands show a similar gap on the questions closest to a purchase, as our e-commerce post sets out, and across all 826 audits from April to October 2026 more than half, at least 51%, were never named (n=826).
Engines read what stays on the page after the sale
Real estate marketing is built for the moment of sale: a launch, a listing, an open house. The engines read what stays on the page between those moments, and an engine can only shortlist a business it can describe.
If the facts that make a builder, a brokerage or a property manager the right choice live in brochures, images and event pages, the engine has nothing to repeat. The question that fails least asks which provider is trusted or best reviewed: at least 67% of 49 times, across 36 audits (Simplee Digital first-party data, doiexist.ai, April to October 2026), and we think reviews help because they are plain text that says who a business is good for.
One page that answers which provider to choose
Write one page that answers "which should I choose" for your own business. Say who you build for or work with, where, at what stage, and what makes you the right call for that buyer.
Keep the facts identical everywhere they appear: your site, your listings and the profiles engines read. Developers can follow the 90-day order we recommend before a condo launch, which starts with exactly this page.
Developers are too few here to report alone
These are real estate businesses of every kind, not only developers, and developers are too few among the 59 to report on their own (n=59, April to October 2026). We are unsure how far the developer figures differ.
Most of these audits ran on our earlier audit pipeline, and stored answers for these audits were cut off at 2,000 characters, so a mention late in a long answer can be missed. The audits also record no country.
How this post counts a skipped question
- A "which" question asks an engine which provider a buyer should choose.
- Failure rate is the share of times a question type was asked and no engine named the audited business.
Source: Simplee Digital first-party data, from doiexist.ai audits (59 audits, April 8 to October 19, 2026). Brand-free recompute (view blog.verify_brandless, group Real estate and property) of the benchmark export, pulled 2026-11-09. The method is on our methodology page. Methodology