Which NYC Realtors Does AI Recommend? 2026 Study

Study date: November 11, 2026

New York City has thousands of agents, but when we asked AI engines 25 everyday questions about finding one, the answers kept coming back to the same few brokerages. Compass, Corcoran, Brown Harris Stevens and Douglas Elliman took 145 of the 287 brokerage and team mentions, or 50.5%.

Compass led with 63 mentions (22.0%), Corcoran had 38, Brown Harris Stevens 23 and Douglas Elliman 21. Sotheby's International Realty was fifth with 13. In all, ChatGPT, Perplexity and Google AI Overview named 35 brokerages and teams.

This is the data the engines returned in one Beket AI audit on one day. It isn't our ranking of who sells homes best, and AI answers change over time.

Key numbers

Measure	Result
Study date	November 11, 2026
Location setting	New York City, New York, United States
Questions	25 unbranded buyer and seller questions, in English
Engines	ChatGPT, Perplexity, Google AI Overview
Answers returned	60 of 75 (ChatGPT 25, Perplexity 25, Google AI Overview 10)
Brokerages and teams named	35
Brokerage and team mentions	287
Named on all three engines	9
Named on only one engine	13
Citation events	505
Most-cited domain	realtor.com, 57 citation events

Who AI named most

Rank	Brokerage or team	Mentions	Named first	ChatGPT	Perplexity	Google AI Overview
1	Compass	63	5	42	17	4
2	Corcoran	38	3	27	2	9
3	Brown Harris Stevens	23	1	17	5	1
4	Douglas Elliman	21	1	14	4	3
5	Sotheby's International Realty	13	4	9	2	2
6	SERHANT.	12	0	8	1	3
7	Hauseit	10	2	5	1	4
8	The Bizzarro Agency	10	2	5	2	3
9	ELIKA Real Estate	7	3	1	3	3
10	Manhattan Realty Group	7	0	6	0	1

The other 25 brokerages and teams shared 83 mentions, between 2 and 7 each. Ties are listed in the market report's order. "Named first" counts answers where that name came first.

The engines mostly named brokerages, not individual agents. A few named teams did show up, such as the Michael J. Franco Team, The Antigua Team and The Bizzarro Agency. We kept teams apart from their parent brokerage, so The Litvak Team @ Compass (4 mentions) is its own row. Folding it into Compass would take Compass from 63 to 67.

Two names on the list work differently from a traditional brokerage. Hauseit (10 mentions) and Prevu (5) are licensed brokerages that advertise buyer rebates or reduced listing commissions. With questions like "low commission real estate agents" in the set, an AI answer can point buyers to a pricing model, not only a big brand.

Nine names every engine returned

Compass, Corcoran, Brown Harris Stevens, Douglas Elliman, Sotheby's International Realty, SERHANT., Hauseit, The Bizzarro Agency and ELIKA Real Estate were named at least once by all three engines. Another 13 showed up on two engines, and 13 on just one.

That is a tighter overlap than we saw for New York marketing agencies, where 12 of 54 agencies made all three engines. Even so, the big names weren't spread evenly across engines, as the next section shows.

Engine by engine

Engine	Questions answered	Names	Mentions	Top names
ChatGPT	25/25	29	188	Compass 42, Corcoran 27, Brown Harris Stevens 17
Perplexity	25/25	19	52	Compass 17, Brown Harris Stevens 5, Douglas Elliman 4
Google AI Overview	10/25	18	47	Corcoran 9, then Compass, Digs Realty Group and Hauseit at 4 each

ChatGPT did most of the naming, with 188 of the 287 mentions (65.5%). It leaned on the big four: Compass, Corcoran, Brown Harris Stevens and Douglas Elliman made up 100 of its 188 mentions.

Perplexity answered every question but named fewer brokerages per answer. Compass dominated there, with 17 of Perplexity's 52 mentions. Perplexity also pointed to listing portals more often than the other engines did, which we cover below.

Google AI Overview answered only 10 of the 25 questions, but it named brokerages densely when it did answer: 47 mentions across 18 names. It was also the only engine where Corcoran beat Compass, 9 to 4. Digs Realty Group and The Glazer Team got all or nearly all their mentions from Google AI Overview, so a ChatGPT-only check would have missed them.

The 15 questions Google AI Overview didn't answer are listed in the method section. We didn't count them as zero mentions.

The sites behind the answers

The engines also named listing portals and agent-matching sites, which we kept out of the brokerage table. StreetEasy was named 20 times and Zillow 16 times, and Perplexity accounted for 19 of the StreetEasy mentions and 9 of Zillow's. HomeLight (9), Clever Real Estate (5) and FastExpert (3) were named too.

The cited sources tell the same story. Of 505 citation events, realtor.com was the most-cited domain with 57. Six listing, review and agent-matching sites in the top ten (realtor.com, StreetEasy, Zillow, Clever, U.S. News Real Estate and Homes.com) added up to 144 citation events, or 28.5% of all citations. Three brokerage or team sites made the top ten: compass.com (26), bhsusa.com (12) and michaeljfranco.com (10).

Domain	Citation events
realtor.com	57
streeteasy.com	27
compass.com	26
zillow.com	19
listwithclever.com	16
realestate.usnews.com	16
google.com	15
bhsusa.com	12
michaeljfranco.com	10
homes.com	9

One detail is worth a look if you work in a neighborhood that shares its name with another town. In the citations we could inspect, one U.S. News agent page cited for "realtors in Williamsburg" was for Williamsburg, Virginia, not Brooklyn. Location settings help, but ambiguous place names can still pull in the wrong market.

For an agent, this means your profiles on realtor.com, StreetEasy, Zillow and similar sites are part of what AI reads about you, alongside your own website and your brokerage's agent page.

New York next to Seattle and Vancouver

We asked a similar set of realtor questions, with local place names, in our Seattle realtors study and our Vancouver realtors study.

City	Study date	Names	Mentions	Top name	Top name's share
Vancouver	Oct 23, 2026	38	237	Oakwyn Realty (36)	15.2%
Seattle	Oct 27, 2026	33	277	Windermere (66)	23.8%
New York City	Nov 11, 2026	35	287	Compass (63)	22.0%

In all three cities, one brokerage led clearly and the rest of the field was spread thin. Treat the comparison carefully: the Seattle and Vancouver runs added 4 discovery prompts each, and Google AI Overview answered a different number of questions in each city, so these aren't equal samples. Our Google Maps vs AI study found that of the 50 businesses in the Google Maps top 3 for 25 Vancouver realtor searches, the AI engines named only 7.

Why it matters for agents

Most buyers still hire someone. In the National Association of Realtors' 2025 Profile of Home Buyers and Sellers, 88% of buyers purchased through an agent or broker and 91% of sellers sold with one. So the real question is which agent a buyer finds, and AI answers are now one of the places that shortlist gets made.

Those shortlists also move. BrightLocal's Local AI Visibility Study of about 200,000 non-branded local prompts found only 20% to 33% average overlap when the same prompt was repeated on the same platform. One run is a snapshot.

What an NYC agent can check

1. Ask the questions buyers ask. Use unbranded questions like the 25 below, not your own name. Our guide to checking whether ChatGPT recommends your business shows how to do it by hand.
2. Check each engine on its own. Google AI Overview favored Corcoran here, while ChatGPT and Perplexity favored Compass.
3. Check your borough and neighborhood questions. Brooklyn, Queens, Harlem and Williamsburg questions can pull in different names, and sometimes the wrong place.
4. Keep your portal profiles current. realtor.com, StreetEasy and Zillow were among the most-cited sources.
5. Re-check over time. Answers changed between repeated runs in BrightLocal's data, so one check isn't enough.

The free Beket AI audit runs 25 prompts across ChatGPT, Perplexity and Google AI Overview for your own brand. Beket AI Pro is $49 a month with daily re-audits, starting at 50 prompts across the same three engines.

Method and caveats

Beket AI ran one market audit for realtors in New York City on November 11, 2026, created at 6:55 PM PT and completed at 7:05 PM PT. We used 25 unbranded buyer and seller questions in English, set the location to New York City, New York, United States, and used three engines: ChatGPT, Perplexity and Google AI Overview. Extra prompt generation was off, so 75 answers were requested. The questions mirror our Vancouver and Seattle realtor studies, with place names changed to New York boroughs and neighborhoods. "Presale condo" became "new development condo," because presale is a Vancouver term.

The final answer log has 60 answers, with no duplicate question and engine pairs and no off-topic rows. ChatGPT and Perplexity answered all 25 questions. Google AI Overview returned no answer for 15 questions: realtors in Williamsburg; best realtors in New York City; realtors for townhouses in Brooklyn; top rated real estate agents with good reviews; condo specialist realtors in downtown Manhattan; commercial real estate agents; how to choose a real estate agent; realtors in Harlem; selling a house on Staten Island; first time home buyers; realtors in the Bronx; realtors in Queens; new development condo realtors; Mandarin-speaking realtors; and realtors in Brooklyn. Missing answers are not counted as zero mentions, and the tools don't show why they're missing.

The raw report had 44 labels and 340 mentions. We removed five labels that aren't brokerages (StreetEasy, Zillow, HomeLight, Clever Real Estate and FastExpert, 53 mentions together) and merged brokerage name variants, such as The Corcoran Group into Corcoran. That left 35 brokerages and teams with 287 mentions. Marcus & Millichap, Cushman & Wakefield and Meridian Retail Leasing are commercial brokerages; we kept them because one question asked about commercial agents. These are editorial calls.

Counts are mention events, not recommendation rates. One answer can name a brokerage more than once, and we didn't infer how many distinct questions named each one. Being named doesn't verify a license, an NYC office or quality of service. The citation domain counts come from the market report's top ten, and the URLs we inspected were a sample of 100.

AI answers change over time, and this is a single run. If your brokerage or team is in this study and you'd like another run, you can ask Beket AI for one.

All 25 questions

1. how to find a New York City realtor
2. realty services vs selling privately in New York City
3. realtors in Brooklyn
4. how much do realtors charge in New York City
5. luxury real estate agents in Manhattan
6. best realtors in New York City
7. low commission real estate agents in New York City
8. top rated real estate agents in New York City with good reviews
9. realtors for townhouses in Brooklyn
10. how to choose a real estate agent in New York City
11. realtors for downsizing seniors in New York City
12. realtors in Queens
13. new development condo realtors in New York City
14. realtors in Williamsburg
15. realtors for selling a house on Staten Island
16. realtors in the Bronx
17. realtors for first time home buyers in New York City
18. commercial real estate agents in New York City
19. real estate teams in New York City
20. realtors in Harlem
21. best listing agents to sell a house in New York City
22. condo specialist realtors in downtown Manhattan
23. real estate agents for investment properties in New York City
24. realtors who speak Mandarin in New York City
25. buyers agent in New York City

Bottom line

In New York City, AI engines mostly hand buyers the same short list of big brokerages, led by Compass and Corcoran, and they lean on realtor.com, StreetEasy and Zillow to build it. Smaller teams can still show up, but usually on one engine and not the others. If you're an NYC agent, check each engine with the questions your clients actually ask.