Which Local Businesses Does AI Recommend? What 9 Studies in Seattle and Vancouver Show (2026)

Between October 16 and October 25, 2026, we ran nine studies with Beket AI. Each one asked AI engines the kind of everyday, unbranded questions people type when they need a local business: "best plumber Seattle," "movers in Burnaby," "family lawyers in Vancouver." None of the questions named a business. Then we counted which businesses the answers named.

This post puts the nine studies side by side. Three are from Seattle (dentists, house cleaning, plumbers) and six are from Vancouver, BC (movers, electricians, lawyers, mortgage brokers, realtors, marketing agencies).

Four patterns held up across them:

• The answers spread wide. Each study surfaced 21 to 49 businesses. The top four names took 28.4% to 49.5% of mentions, so in every study more than half of all mentions went to businesses outside the top four.
• The engines rarely agreed. In the six Vancouver studies, the share of businesses named by every engine we tested ranged from 0 of 29 (movers) to 11 of 38 (realtors). Only one of the six, realtors, had the same top name on every engine.
• Google AI Overview answered least. Across the six Vancouver studies, the audits recorded a Google AI Overview for 38 of 174 logged queries (21.8%). Perplexity returned an answer for all 174.
• Perplexity cited the most sources in all six Vancouver studies (388 to 484 per study), while ChatGPT gave the most mentions in all five where we tested it.

These are the answers the engines returned on those dates. AI answers change, so treat every number here as a snapshot, not a verdict. They are not Beket AI's opinion and not a ranking of quality.

Contents

1. The nine studies at a glance
2. How many businesses AI named
3. How concentrated the answers were
4. How often the engines agreed
5. Which engine answered most and least
6. Not every top name was a local business
7. What local businesses can take from nine studies
8. How the studies differ, and the limits
9. Request another run

The nine studies at a glance

Each row links to the full study, with its methodology, every question we asked and the complete results.

Study	City	Study date	Questions	Engines	Businesses named	Mentions	Most-named	Top-four share	Named by every engine tested
Seattle dentists study	Seattle	Oct 16, 2026	24	ChatGPT, Google AI Overview, Perplexity (combined)	35 clinics (39 names)	165 clinic (205 total)	Seattle Dental Co (clinic)	31.5%	Not reported
Seattle house cleaning study	Seattle	Oct 16, 2026	22	ChatGPT, Google AI Overview, Perplexity (combined)	25	313	Simply Clean	49.5%	Not reported
Seattle plumbers study	Seattle	Oct 16, 2026	21	ChatGPT, Google AI Overview, Perplexity (combined)	23	143	Ben’s Plumbing & Drain Cleaning and Beacon Plumbing (tie)	44.8%	Not reported
Vancouver movers study	Vancouver	Oct 20, 2026	25 per engine (23 shared)	ChatGPT, Google AI Overview, Perplexity (separate runs)	29	219	BOSS Moving	46.1%	0 of 29
Vancouver electricians study	Vancouver	Oct 21, 2026	25	Google AI Overview, Perplexity (separate runs)	21	112	Kato Electrical	42.0%	6 of 21
Vancouver lawyers study	Vancouver	Oct 23, 2026	25 + 4 discovery	ChatGPT, Google AI Overview, Perplexity (one audit)	39	210	Fasken	28.6%	1 of 39
Vancouver mortgage brokers study	Vancouver	Oct 23, 2026	25 + 4 discovery	ChatGPT, Google AI Overview, Perplexity (one audit)	35	236	Pinsky Mortgages	28.4%	6 of 35
Vancouver realtors study	Vancouver	Oct 23, 2026	25 + 4 discovery	ChatGPT, Google AI Overview, Perplexity (one audit)	38	237	Oakwyn Realty	34.2%	11 of 38
Vancouver marketing agencies study	Vancouver	Oct 25, 2026	25 + 4 discovery	ChatGPT, Google AI Overview, Perplexity (one audit)	49	422	Major Tom	32.0%	3 of 49

Source: Beket AI audits, October 16–25, 2026, as published in each study. "Businesses named" counts the business type the study covered after combining name variants (the dentists study used exact names). Dentists: clinics only (Delta Dental, an insurer, had more mentions overall); the top-four share is of the 165 clinic mentions. House cleaning: the 25 names include one review site. Plumbers: one directory left out. "Top-four share" is the top four names' mentions divided by all mentions in that study's main count; where names tied at #4, the tied count is the same either way. "Named by every engine tested" is available only where results were split by engine; for electricians it means both engines.

How many businesses AI named

Across the nine studies, the counts include 293 businesses of the types we studied (plus one review site kept in the house cleaning count), between 21 (Vancouver electricians) and 49 (Vancouver marketing agencies). The median study named 35.

Source: Beket AI audits, October 16–25, 2026.

Two things stand out:

• Professional services drew the longest lists. The four Vancouver studies run on October 23 and 25 named 35 to 49 businesses each, against 21 to 29 for the trades (plumbers, movers, electricians). Marketing agencies, which mostly sell to other businesses, drew the most of both: 49 agencies and 422 mentions.
• Most names came up only a few times. In the Seattle dentists study, 27 of the 39 names were mentioned 4 times or fewer. In the Seattle plumbers study, 11 plumbing businesses were mentioned 3 times or fewer.

Some of the gap between studies comes from method, not market. The October 23 and 25 studies counted 4 extra discovery prompts on ChatGPT and Perplexity, and the Seattle dentists study did not combine name variants. Compare the studies by pattern, not by exact count.

How concentrated the answers were

No single business dominated any of the nine studies.

• The most-named business took 7.6% to 19.2% of mentions. The low was Fasken among Vancouver law firms (16 of 210). The high was Simply Clean among Seattle house cleaners (60 of 313).
• The top four took less than half everywhere. Seattle house cleaning was the most concentrated, at 49.5% (155 of 313). Vancouver mortgage brokers (28.4%, 67 of 236) and lawyers (28.6%, 60 of 210) were the least.
• Being named most is not the same as being named first. In Seattle, The Dazzle Cleaning Company was named first 11 times against 7 for Simply Clean, the most-mentioned name. Beacon Plumbing had 17 mentions and was never named first. Kerrisdale Electric led Google AI Overview's electrician mentions (9) and was never named first there.

For a local business, that spread cuts both ways. There is no locked-in winner to displace. But a business outside the top few may come up only when someone asks a narrow question, such as a neighbourhood or a specific job.

How often the engines agreed

Six of the studies, all in Vancouver, split results by engine. The three Seattle studies added the engines together, so they can't answer this question.

Study	Engines tested	Named by every engine	Named by only one engine	Same top name on every engine?
Movers	3	0 of 29	24 of 29	No
Electricians	2	6 of 21	15 of 21	No
Lawyers	3	1 of 39	11 of 39	No
Mortgage brokers	3	6 of 35	11 of 35	No (two of three)
Realtors	3	11 of 38	9 of 38	Yes: Oakwyn Realty
Marketing agencies	3	3 of 49	21 of 49	No

Source: Beket AI audits, October 20–25, 2026, as published in each study.

The clearest splits:

• Movers: BOSS Moving led ChatGPT with 34 mentions. Ferguson Moving & Storage led Perplexity with 11 and was not in ChatGPT's results at all.
• Electricians: Kerrisdale Electric, Google AI Overview's most-mentioned electrician, was not in Perplexity's results. Kato Electrical, Perplexity's leader with 14 mentions, got 2 on Google AI Overview.
• Lawyers: ChatGPT named Blakes most (13), Perplexity named Watson Goepel LLP most (8) and Google AI Overview named YLaw most (3). Fasken led the combined count without leading any single engine.
• Marketing agencies: Takt led ChatGPT with 34 mentions but got 5 on Perplexity. Major Tom led Perplexity with 24.

The only study where every engine agreed on the top name was realtors: Oakwyn Realty led ChatGPT (17), Perplexity (10) and Google AI Overview (9).

Read the "every engine" column with one caveat: Google AI Overview named very few businesses in most studies (2 law firms, 5 agencies), which caps how many businesses all three engines could share. That is why we've argued you can't trust one AI engine to tell you how visible a business is.

Which engine answered most and least

Each Vancouver run logged 29 queries per engine: the questions plus discovery queries the audit adds on its own.

Source: Beket AI audits, October 20–25, 2026.

• Google AI Overview: an answer was recorded for 38 of 174 queries (21.8%), from 3 of 29 for lawyers to 12 of 29 for electricians. For the rest, the audits recorded no AI Overview.
• Perplexity: an answer for all 29 queries in all six studies. It also cited the most sources every time: 412 (movers), 388 (electricians), 389 (lawyers), 465 (mortgage brokers), 416 (realtors) and 484 (marketing agencies).
• ChatGPT: tested in five of the six studies, with an answer for all 29 queries each time. It gave the most business mentions in all five, from 46.0% of the total (realtors, 109 of 237) to 72.0% (marketing agencies, 304 of 422). It cited fewer sources than Perplexity, 92 to 181 per study.

So Perplexity read the most, ChatGPT named the most, and Google AI Overview often did not answer at all for these local questions. Google's low rate here fits what we found when we looked at what percentage of Google searches have AI Overviews: the wording of a local search changes whether an AI Overview shows. For how ChatGPT finds and cites websites when it searches, see how ChatGPT search works.

Not every top name was a local business

In two of the nine studies, the name that came up most in at least one view was not the type of business we studied.

• Seattle dentists: the most-mentioned name overall was Delta Dental, an insurer, with 29 mentions (14.1%). One question asked for a dentist that takes Delta Dental.
• Vancouver electricians: on Google AI Overview, BC Hydro, the provincial utility, had 10 mentions, more than any electrician. Several questions were about EV chargers, panel upgrades and heat pumps.

Other studies saw the same thing on a smaller scale. A public auto insurer came up 7 times in the lawyers study, banks and a credit union appeared among mortgage options, and review sites or directories turned up in the dentists, house cleaning, plumbers and movers studies. For a local business, the competition for an AI answer includes utilities, insurers and directories, not just rivals.

Some answers also crossed the border. In the movers study, 12 of the 100 Perplexity citations we could export pointed to pages about movers in Vancouver, Washington. We've written more about that in does AI search even know your business is Canadian?

What local businesses can take from nine studies

• Check more than one engine. In five of six Vancouver studies, the engines' top names differed. A business can lead on one engine and be absent on another.
• Don't read Google AI Overview's silence as a result. For most of these local queries, no AI Overview was recorded. Check the engines that did answer, too.
• Answer the narrow questions. The spread of names means specific questions, such as a neighbourhood, a service or a type of client, surface different businesses. Give each service and area its own clear page.
• Keep your name the same everywhere. Every study except dentists had to combine spellings of the same business, from "BOSS Moving" and "BOSS Moving Vancouver" to "Blakes" and "Blake, Cassels & Graydon LLP."
• Say where you are. Put your city, province or state, and service areas where they're easy to read, especially if your city's name exists elsewhere.
• Check again. These studies are snapshots. One check is a snapshot, not a verdict.

How the studies differ, and the limits

• Different dates. The three Seattle studies ran on October 16, 2026; Vancouver movers on October 20; electricians on October 21; lawyers, mortgage brokers and realtors on October 23; and marketing agencies on October 25. These are the answers the engines returned on those dates. AI answers change, and a run on another day could differ.
• Different setups. The Seattle studies were one combined audit per industry, so they don't split results by engine. Movers and electricians used one run per engine. The October 23 and 25 studies used one audit with a per-engine split and 4 discovery prompts. The electricians study tested two engines, not three.
• Different counting rules. The dentists study used exact names; the others combined clear variants. Each study set aside different non-business names. Every name in the audit results had at least 2 mentions, so businesses named once may be missing.
• Small numbers. One or two mentions can change a rank, especially on Google AI Overview.
• Mentions, not quality. A mention means an engine gave a name. It says nothing about service, licensing or price. We did not rate, verify or rank any business ourselves.

Each study's own post lists every question we asked and its full method.

Request another run

AI answers move, and we plan to repeat these studies to see what changes. If you run a business in one of these industries, or in another city, you can request another run, or your own audit, at beket.ai.

A free Beket AI audit covers 25 prompts across 3 engines. Beket AI Pro is $49/month and re-audits daily, starting at 50 prompts across 3 engines.