Does Google AI Mode Show Ads? 42 of 100 Local Hiring Answers Had One.

Update, October 29, 2026: I’m re-running this test, and the first repeat found far fewer ads. On October 29 I asked the same locked questions again, with each search set to its own city instead of run from my location near Spokane. So far 4 of 75 answers had a sponsored block (5.3%, 95% CI 2.1% to 12.9%), against 42 of 100 in the original run. Spokane went from 12 of 20 to 2 of 20. The gap is too big to be chance alone, but I don’t know the cause yet. Three things changed between the runs: the location setting, the day, and the AI Mode page itself, which now usually loads without the “may include mistakes” footer. In one answer the sponsored block sat after the follow-up questions. Until the repeats are done, read the figures below as one run on October 27, not a typical rate. A third run is set for October 31, and I’ll rewrite this post with all three runs after it.

Does Google AI Mode show ads in local answers?

Yes. On October 27, 2026, I asked Google’s AI Mode 100 “who should I hire” questions, the same 20 questions in each of five U.S. cities: Spokane, Seattle, Denver, Dallas and Kalispell, Montana. 42 of the 100 answers had a block labeled “Sponsored” inside the answer. Home services drew ads in 34 of 55 answers. Professional services, like lawyers, dentists and CPAs, drew one in 30.

If you’re a plumber, electrician, exterminator, house cleaner or mover, that means a competitor can pay to sit inside the answer when someone asks the AI who to hire.

How I tested it

I wrote 20 questions a local buyer would ask, like “Who is the best plumber in Denver, CO?” or “Best CPA for a small business in Dallas.” Eleven are home services. Six are professional services: a family dentist, a personal injury lawyer, a real estate attorney, a CPA, a chiropractor and a marketing agency. Three are auto repair, a vet and a furniture store. The furniture question is a shopping question more than a hiring one, so I grouped it with auto repair and vets instead of home services.

Each city got the same list with the city name swapped. The one exception is Spokane’s electrician question, which named Spokane Valley. I saved each city’s list and its SHA-256 hash before that city’s first question ran, and I haven’t changed them since. The lists and hashes are at the end of this post.

I asked each question once, signed out, in a desktop browser, from my own location near Spokane. I counted an answer as having an ad when a block labeled “Sponsored” appeared in the answer text before Google’s “AI responses may include mistakes” footer. I read the text around all 42 hits to confirm each was an advertiser block. I also searched the other 58 answers for any other ad label and found none.

I started with Spokane, where I live, and Kalispell, a small city in northwest Montana. Then I added Seattle, Denver and Dallas: three larger markets, two of them outside the Northwest.

How often did AI Mode answers include an ad?

City	Answers with an ad	95% confidence interval
Spokane, WA	12 of 20 (60%)	38.7% to 78.1%
Seattle, WA	9 of 20 (45%)	25.8% to 65.8%
Denver, CO	7 of 20 (35%)	18.1% to 56.7%
Dallas, TX	7 of 20 (35%)	18.1% to 56.7%
Kalispell, MT	7 of 20 (35%)	18.1% to 56.7%
All five	42 of 100 (42%)	32.8% to 51.8%

Every city’s interval overlaps every other’s, so I wouldn’t read anything into the city order. Spokane also isn’t a fair comparison. It’s where I searched from, and it’s the only city I ran in the morning, so local advertisers could have reached me by location as well as by the city in the question.

The intervals treat each answer as independent. They aren’t fully: the same 20 questions repeat in each city, and many came out the same everywhere. The real uncertainty is wider than the table shows, and it doesn’t account for answers changing from one run to the next.

Which local businesses get ads in AI Mode answers?

The type of business mattered much more than the city.

Business type	Answers with an ad	95% confidence interval
Home services (11 questions per city)	34 of 55 (62%)	48.6% to 73.5%
Auto repair, vet, furniture store (3 per city)	7 of 15 (47%)	24.8% to 69.9%
Professional services (6 per city)	1 of 30 (3%)	0.6% to 16.7%

By city, home services drew ads in 9 of 11 questions in Spokane, 7 of 11 in Seattle and 6 of 11 in Denver, Dallas and Kalispell.

By question, here’s how many of the five cities had an ad in the answer:

Cities where the AI Mode answer had an ad, by question

Out of 5 cities, one run each, October 27, 2026

Plumber

5

Electrician

5

Pest control

5

House cleaning

5

Moving company

5

Roofing

4

Furniture store

4

Window replacement

2

Veterinarian

2

HVAC repair

1

Landscaping

1

Kitchen remodel

1

Auto repair

1

Real estate attorney

1

Concrete work

0

Family dentist

0

Personal injury lawyer

0

CPA

0

Chiropractor

0

Marketing agency

0

012345

Question	Cities with an ad (of 5)
Plumber	5
Electrician	5
Pest control	5
House cleaning	5
Moving company	5
Roofing	4
Furniture store	4
Window replacement	2
Veterinarian	2
HVAC repair	1
Landscaping	1
Kitchen remodel	1
Auto repair	1
Real estate attorney	1
Concrete work	0
Family dentist	0
Personal injury lawyer	0
CPA	0
Chiropractor	0
Marketing agency	0

The only professional-services ad in the whole test was a law firm’s text ad on the Dallas real estate attorney question.

How does this compare to other AI Mode ad studies?

The biggest study I know of is SE Ranking’s, published July 14, 2026. They checked 50,032 U.S. commercial keywords on June 30 and found AI Mode text ads on 29.45% of them. Healthcare had the lowest ad presence in their data, 2.64%, and the low-ad niches leaned toward informational and “your money or your life” topics. That lines up with dentists, chiropractors, lawyers and CPAs drawing almost nothing here.

The two numbers aren’t directly comparable. Their sample is 500 times bigger and covers a broad mix of commercial searches. Mine covers one narrow, high-intent kind: asking who to hire in a named city. They also counted text ads only. Counting just the text ads here, 14 of my 100 answers had one. Most of the ads I saw were business listings.

What do the ads look like?

28 of the 42 ads were business cards or short business lists, usually under a line like “To get a quote for your project, you can connect with these local plumbers.” One, in Seattle, was a national furniture retailer rather than a local business. The other 14 were single text ads with the advertiser’s name, a line of ad copy and their website. I sorted them by reading each one.

In all 42, the “Sponsored” label sat in the last 40% of the answer text, after most of the AI’s own answer. No answer had more than one sponsored block. I counted that from the saved answer text, which I’m not publishing because it names the advertisers.

I’m leaving the businesses in the answers unnamed. The point is how Google builds the answer.

Are AI Mode’s recommendations themselves sponsored?

The paid placements I saw were labeled and sat in their own block. I didn’t measure whether the advertisers also showed up in the AI’s unpaid picks, so I can’t say whether paying for the ad has any effect on those.

What this means if you own a local business

1. The AI answer now has paid space in it. For plumbing, electrical, pest control, cleaning and moving questions, my one run drew an ad in all five cities. It’s labeled, but it sits inside the answer a buyer is reading. If you already pay for those clicks on regular search, my home services cost-per-lead benchmarks show what plumbing and roofing leads cost there.
2. If you’re in professional services, there was almost no paid space to compete with. Dentist, CPA and chiropractor questions drew no ads in any city, and lawyer questions drew one in ten. That leaves the unpaid picks. In my six-audit study, the businesses engines named were the ones directories, reviews and roundups vouched for.
3. Ask whoever runs your Google Ads whether you’re eligible to show here, and whether they can see impressions from it. Google has changed which campaign types can serve in AI Mode more than once, so check the current rules against your own setup.

Full disclosure: I run Google Ads for clients, so I have a stake in this.

What I don’t know

• One run per question. AI answers change from one run to the next, and so do ads. A second run the same day could land on different numbers.
• One window of time. All 100 questions ran between about 11 a.m. and 2 p.m. Pacific. I didn’t test whether time of day matters.
• One engine. I only checked Google’s AI Mode. ChatGPT, Gemini and Perplexity weren’t part of this.
• One location. Someone searching from a phone in Dallas could see something different.
• Small samples. 20 questions per city is enough to see the pattern by business type, not to rank cities.

The data

• All 100 answers, one row each: city, question, ad or no ad, ad format and where the label sat. Advertiser names are left out.
• The five question lists, with the SHA-256 hash I saved before each city ran:
  • Spokane: 4560206cbfbfcdb9d26977d231a9340906eb7328fa449bb583d89e01bb8ea9f4
  • Kalispell: d3d34f5ce3a40d8e56f13998ea90efebc25bf096a079f040549f30916faf4caa
  • Seattle: 58dfef61cc40b0e6a3626c66523d94c910f5c671eb1fb576735fa077fb327d1c
  • Denver: 7aeaa05e31e6d8514cb865d30606c2f6bcbae555d843157fd30f826a974893df
  • Dallas: 4586871f8be7273c34c43de099c50839cc37155124f73e20420cfe28b9255c73

If you want to know what AI assistants say when someone asks who to hire in your category, I put together a free snapshot for that on the AI answer audit page.

For how often AI answers name a business at all, and how often they get its facts wrong, see my AI answers field report. The free What AI says about you template lets you run the questions on your own business.

See how I run Google Ads for service businesses.