12 home-buyer prompts without any developer's name, run daily in India on ChatGPT, Gemini, Perplexity and Google AI Overviews for two residential developers (one in Bengaluru, one across Bengaluru, Chennai and Pune): 64 runs, 1,617 answers and 16,890 citation sightings, plus two live 16-query Google AI Overview panels on 2026-08-17 and 2026-09-15.
Key findings
| Finding | What it measures | What it means |
|---|---|---|
| 0 of 407 | Answers naming the tracked developer on generic buyer-research questions, all four engines, 2026-08-17 to 2026-09-28 | Questions about choosing a property, amenities and types of development never named the tracked developers, and 83.3% of these answers named no developer at all. Other developers' websites still supplied 45.9% of their citations. |
| 59% | Share of citations to developer websites that point at blog or guide pages, 2026-08-30 to 2026-09-28 | 2,051 of 3,473 citations to developer sites landed on blog, guide or tips URLs rather than project pages. Engines use developers as publishers of buyer advice, often without naming them. |
| 31.2% | Listing portals' share of all 16,890 citation sightings, 2026-08-17 to 2026-09-28 | Portals and broker project sites are the largest single source. On budget queries for a specific locality they supply 69.5% of citations, against 1.6% for developer sites. |
| 98.9% | 'Best developers in [city]' answers naming at least one of the same four large developers, all four engines, 2026-08-17 to 2026-09-28 | 174 of 176 answers named one of four large developers and listed 6.9 developers on average. The tracked mid-size developer appeared in 7 of 176 (4.0%). |
| 0.2% | Tracked developers' own sites as a share of all 16,890 citation sightings, 2026-08-17 to 2026-09-28 | Their own websites were almost never a source. When one developer's site was cited (25 answers, 2026-09-09 to 2026-09-28), the developer was named in only 2 of those answers. |
| 2 of 25 vs 5 of 7 | Live Google AI Overviews mentioning the developer, unbranded vs branded queries, India, 2026-08-17 and 2026-09-15 | Adding the developer's name to the query made it appear in most answers; without it, the developer appeared in 2 of 25, both on locality questions where it had a project page. |
Where the answers come from
Share of all citation sightings by source type, all four engines, 2026-08-17 to 2026-09-28.
Listing portals31.2%
Other developers' sites26.1%
Other (Google Maps and search links, image hosts, documents)20.7%
Blogs and guides10.9%
News and media4.8%
Government and regulator3.0%
Review and marketplace sites1.9%
Reddit0.6%
YouTube0.4%
The tracked developer's own site0.2%
Wikipedia0.1%
Each engine answers differently
- Gemini leans on developer websites: 54.2% of its 2,958 citation sightings went to developers' sites, against 6.2% for ChatGPT, and it named the tracked developers most often (47 of 403 answers, 11.7%), 2026-08-17 to 2026-09-28.
- Perplexity reads the widest set of sources, 16.5 citations per answer, with 40.0% going to listing portals, yet only 31.5% of its answers named any developer, 2026-08-17 to 2026-09-28.
- ChatGPT is the only engine that leans on regulators and planning bodies such as RERA and municipal sites (12.7% of its citations) and named the tracked developers in 24 of 408 answers (5.9%), 2026-08-17 to 2026-09-28.
- Google AI Overviews named at least one developer in 69.2% of answers, the most of any engine, but 55.9% of its cited links were Google Maps or search pages with no attributable publisher, and a scraper outage degraded its runs from 2026-09-12 to 2026-09-28.
What wins each type of buyer question
| Question type | Example | What gets named and cited |
|---|---|---|
| Locality + budget | 3 BHK in [locality] within [budget] | Listing portals supply 69.5% of citations and developer sites 1.6%, yet a developer with a project in that locality was named in 21.6% of answers (76 of 352), 2026-08-17 to 2026-09-28. |
| Upcoming projects by locality | Upcoming residential projects in prime localities of [city] | Answers are lists: 97.4% named a developer (6.4 on average), 96.1% included one of four large developers, and portals supplied 51.3% of citations (77 answers, 2026-09-09 to 2026-09-28). |
| Best developers in a city | Best residential developers in [city] for family homes | Four large developers appear in 98.9% of answers, with portals (32.2%) and other developers' sites (24.6%) as the main sources; the tracked mid-size developer made 4.0%, 2026-08-17 to 2026-09-28. |
| Trust and verification | How to find trusted residential developers in [city] | 81.8% of answers named a developer and the tracked one reached 17.0%; review and marketplace sites (5.5%) and regulators (3.9%) carry more weight here than on any other question type, 2026-08-17 to 2026-09-28. |
| Generic buyer research | Key factors when choosing a property in [city] / what amenities matter in a new development | Other developers' blog posts win the citations (45.9%) but 83.3% of answers name no developer, and the tracked developers were named in 0 of 407, 2026-08-17 to 2026-09-28. |
| Investment | Best areas in [city] for long-term real estate investment | Blogs and guides (26.8%), portals (26.0%) and developer sites (21.4%) split the citations, and the tracked developer was named in 10 of 253 answers (4.0%), 2026-08-17 to 2026-09-28. |
What a developer's marketing team can do
- Build a page for each project, locality and budget combination (for example [BHK] in [locality] under [budget]) with price band, RERA number and possession date, and mirror the same facts on portal listings. These queries named a developer with a local project in 21.6% of answers, and portals supply 69.5% of their citations.
- Write buyer guides so your projects and your name sit inside the answer text, not just in the byline. Generic research answers cited developer sites 45.9% of the time but named the tracked developers 0 of 407 times, and one developer's site was cited in 25 answers while it was named in 2.
- Earn places on third-party 'best developers in [city]' lists, portal developer pages and press roundups. Four large developers hold 98.9% of those answers, and a mid-size developer appeared in 4.0%.
- Publish identity data engines can parse: valid Organization and project schema, one name per project across the site, the RERA record and portals. The tracked developers' own sites drew 0.2% of citations, and an August check found one developer's project schema broken by a single JSON error.
- Keep regulator and review records current. ChatGPT takes 12.7% of its citations from regulators and planning bodies, and trust questions draw 5.5% of citations from review and marketplace sites.
- Measure unbranded questions per engine, every week. Branded queries flatter (5 of 7 live answers vs 2 of 25 unbranded), and one developer's mention rate on ChatGPT, Gemini and Perplexity halved from 12.6% (2026-08-17 to 2026-08-30) to 6.1% (2026-09-15 to 2026-09-28) while its Perplexity rate rose.
The brand studies behind it
- How a Bengaluru developer gets named when buyers ask AI about new launches (Residential developer, Bengaluru)
- How a real estate software vendor gets named in 58% of Perplexity answers (Transaction software, US real estate)
- Where a multi-city developer wins AI answers: local budget questions (Residential developer, 3 Indian cities)
- From zero to named in 13 of 14 Perplexity answers, and how to keep it (Leasing AI, property operators)
Methodology
AirPulse ran each prompt once a day on ChatGPT, Gemini, Perplexity and Google AI Overviews with India as the region, storing every answer and cited URL, from 2026-09-09 to 2026-09-28 for one developer and 2026-08-17 to 2026-09-28 for the other (its first three days, run without an India region, are excluded); the two live panels were single logged-out Google runs from India. A developer counts as named when its name, or a project carrying its name, appears in the answer text, and sources are grouped by hand-assigned domain category, with broker and channel-partner project sites counted as listing portals. A Google AI Overview scraper outage from 2026-09-12 to 2026-09-28 reduced the citations Google AI returned, so changes over time are measured on ChatGPT, Gemini and Perplexity only. Developer, project, competitor and portal names and domains are removed throughout.
For how AirPulse measures AI answers in general, see the measurement methodology. For what we do for developers, see AI visibility for real estate developers.