technicalmarketing.ai Free AI audit

Report

How does DataForSEO's Search Insider 2026 compare with local AI answers?

DataForSEO† (search data API) says its Search Insider 2026 holds 137.9M Google AI Overviews answers1. We asked three assistants 199 local questions and read 597 answers2. Both see Reddit near the top of cited sites, and DataForSEO† reports it second-most-cited2,3. The rest of the lists differ because our questions name a profession and a city.

What did DataForSEO and we each measure?

We are a paying customer of DataForSEO's API, and our AI answers and search volumes were collected through it. This page sets findings side by side and ranks neither study.

Our first study put 199 local questions to ChatGPT, Gemini and Perplexity across 16 US cities and 23 kinds of business on 2026-11-10, with web search on, and collected 597 answers2. Each question took the form "Who is a good dentist in Atlanta, GA?" with the city and the kind of business changed2. It asked each question once, so it is a baseline2. Our repeat study asked 63 questions five times each on 2026-11-11, which gave 945 answers about four professions4. Ours is small: 597 answers in the first study, where DataForSEO† reports 137.9M answers in its AI Overviews corpus1.

Table 1. What each study measured

DataForSEO†Our studies
Sample190M+ keywords, 165M+ AI answers5199 questions, 597 answers; repeat study 945 answers2,4
Answer corpora27.46M ChatGPT answers, 137.9M AI Overviews answers6ChatGPT, Gemini and Perplexity, one answer each per question2
DatesAI Overviews Aug 2025 to Sep 2026; ChatGPT Jul 2025 to Sep 202612026-11-10; repeat 2026-11-112,4
MarketUS, English116 US cities, 23 kinds of business2
Kind of question27.59M ChatGPT prompts, 139.0M AI Overviews prompts; wording NF1Who is a good professional in a city2
MethodSeven datasets collected by DataForSEO7API, web search on, names matched by spelling2

DataForSEO† says it draws on seven datasets that it collects and maintains, and reports them separately because they differ in coverage, observation periods and selection criteria1,7. The wording of the prompts behind its answer corpora is not described in the evidence we hold. Table 2 lists the DataForSEO† findings used below with the page of its report where each sits.

Table 2. DataForSEO† findings used on this page

FindingReport page
ChatGPT and AI Overviews answer corpora; share of answers with no citation6p. 5
Most-cited domains in AI Overviews8p. 20
Most-cited domains in ChatGPT; YouTube in both9p. 21
Reddit's rank and presence in Google's top 103,10pp. 22 and 28
First and fourth priorities3,11pp. 28 and 29
AI Overview rate by industry12,13pp. 42 and 43
AI Overview rate by cost per click14p. 15
AI Overview presence over time; Local Packs15p. 12
Estimated click share15,16pp. 11 and 12
Desktop click-through17p. 10
Referring domains and citations18p. 25

Where do DataForSEO and our data agree?

Platforms favor different sources. DataForSEO† reports that YouTube ranks first in Google AI Overviews with 51.9M citations and 19th in ChatGPT with 211K9. Its fourth priority is to assess SEO and GEO competition separately, because strength on one AI platform does not guarantee visibility on another, it says11. In our data the three assistants named 2,975 businesses and only 122 of them (4.1%) were named by all three2, and 88% of the websites cited were cited by only one assistant2. The units differ: DataForSEO† compares domains cited by two platforms, and we compare businesses and websites across three assistants for local questions. It also says its two top-30 lists are ranked separately for each platform and cannot be compared on absolute citation volume19. Our repeat study adds that two answers from the same assistant shared 43.2% of the businesses named, and two answers from different assistants shared 9.7%4.

Sites beyond your own website matter. DataForSEO† says its first priority is to look beyond your own website, because Reddit threads, YouTube videos, publishers and reference sites also compete for visibility3, and it reports Yelp at 15th among the most-cited domains in AI Overviews with 2.2M citations20. In our data the 25 most-cited websites drew 29% of all citations, and they were directories, review platforms, list sites, publishers and a forum2. The two sets of figures measure different things, since one counts citations over a corpus and the other counts citations over 597 answers. Websites cited in only one city, where a local business's own site would fall, still drew 39% of our citations2.

Reddit is prominent. DataForSEO† says Reddit is the second-most-cited domain in both Google AI Overviews and ChatGPT in its datasets3, and it reports that ChatGPT cites Wikipedia in 4.88M unique answers and Reddit in 3.25M9. Reddit was the most-cited website in our study, in 87 answers2. The counts are of different things: citations and unique answers in its corpora, and answers among our 597 that cited the site.

Where do they differ, and why?

Which sites lead. DataForSEO† reports YouTube as the most-cited domain in AI Overviews, ahead of Reddit, Facebook, Wikipedia and Google, and says six of the ten most-cited domains are social, video or community platforms8. The websites cited in the most answers in our study were led by reddit.com in 87 answers, threebestrated.com in 76 and consumeraffairs.com in 712. YouTube, Facebook and Wikipedia are not among our ten. Three reasons apply. Our questions name a profession and a city and ask for specific businesses2, while DataForSEO† reports 139.0M prompts in its AI Overviews corpus without describing their wording1. The platforms differ: Google AI Overviews and ChatGPT are its two, and ours are ChatGPT, Gemini and Perplexity. The dates differ too, since its lists run through September 2026 and our answers were collected on 2026-11-10.

Reddit also splits by assistant in our data. Gemini cited it in 42% of its answers, ChatGPT in 2% and Perplexity in 0%2. DataForSEO† reports Reddit second in ChatGPT across a corpus that spans GPT-4o through GPT-5.61,3, while our ChatGPT figure is for one model, gpt-6-luna2. The models, the questions and the unit of count all differ.

Answers with no citation. DataForSEO† reports that 27.3% of ChatGPT answers carry no source citation, against 0.7% of Google AI Overview answers6. Our keyed figures are averages: ChatGPT cited 3.5 websites per answer, Gemini 6.5 and Perplexity 7.72. An average cannot be turned into a share of answers with no cited website, so that share is NF in our evidence. We collected our answers with web search on2, and whether the ChatGPT prompts in the DataForSEO† corpus ran with search is NF in the evidence we hold.

AI Overview rate by industry. DataForSEO† reports an overall AI Overview trigger rate of 56.7%21, with Health at 69.68%12, Finance at 68.80%13, Law & Government at 61.07%13, Home & Garden at 49.28%13 and Real Estate at 22.27%13. It says it groups pages by its own industry classification13. Table 3 sets the nearest kind of business from our study beside each industry. The pairing is ours.

Table 3. AI Overview rate by DataForSEO† industry, beside our nearest kind of business

DataForSEO† industryAI Overview rateNearest kind of business in our studyBusinesses named by only one assistant
Health69.68%12Dentists88%2
Finance68.80%13Financial advisors87.5%2
Law & Government61.07%13Family law attorneys75.8%2
Home & Garden49.28%13Roofers70.4%2
Real Estate22.27%13Real estate agents83%2

The columns measure different things: how often Google shows an AI Overview for a keyword, and how often only one assistant names a business. We infer no link between them. We did not measure how often Google shows an AI Overview for a profession, so that rate is NF for our professions.

Costly keywords. DataForSEO† reports that keywords costing $5 or more per click trigger AI Overviews 49.55% of the time, close to the 50.54% for keywords costing $0 to $0.2514. Our cost-per-click data answers a different question, what a click costs: Google Ads data for November 2026 puts "lawyer seo" at $108.78, "dental marketing" at $47.30 and "accountant marketing" at $24.9722. All three are above $5, so they sit in the band DataForSEO† describes, but its rate for those particular searches is NF.

What did only one of the studies measure?

DataForSEO† measured how search results themselves changed. It reports that the share of analyzed Google results pages with an AI Overview rose from 5.93% to 61.43% between September 2024 and August 202615, and that Local Packs reached 25.99% of results pages from 19.98% over the same panel15. It reports an estimate that AI Overview citations capture 42% to 45% of clicks when present, and that the combined click share of positions 2 to 10 falls by about 36% to 42%15,16, and it states that these are modeled estimates and not observed clicks23. In anonymized data from 2K+ websites, it reports desktop click-through for position 1 falling from 17.83% to 7.43% between September 2025 and August 2026, and says its data does not establish what drives the decline16,17. It also reports a Spearman correlation of 0.406 between referring domains and Google AI Overview citations and 0.377 for ChatGPT, and states that these are associations that do not establish that backlinks influence citations18,24.

We measured what the report's figures in our evidence do not contain: which businesses an assistant names, and how much that list varies. In our repeat study an assistant named 6.2 businesses in its first answer and 13.5 different businesses across five answers to the same question4, and a business named in one answer was in another answer from the same assistant 55% of the time4. The share of named businesses that all three assistants named ranged from 1.9% in Charlotte to 7.9% in Seattle2, and from 1.1% for dentists to 6.9% for roofers2.

Both studies looked at links, in different ways. Of 12 contractor websites ranking in Google in Thurston County, Washington, 11 showed between 2,591 and 2,751 referring domains each, almost all spam, and after filtering the median site had about 22 clean referring domains25. DataForSEO† reports a median of 111 referring domains for websites ranking first against 85 for positions 4 to 10, in a sample that excludes large platform domains18,26. The samples and the filtering differ, so we compare no figure with another.

What does neither study show?

How often assistants name the same local business, as a DataForSEO† measure, is NF; the nearest evidence is its domain-level lists, which it says are ranked separately for each platform19. The AI Overview rate for a search that names a profession and a city is NF in both studies; the nearest evidence is its report of a Health rate of 69.68%12. Whether being cited produces a call, a booking or revenue is NF; the nearest evidence is the click shares DataForSEO† reports, which it states are modeled estimates23. Whether links from user-generated sites cause more citations is NF; DataForSEO† states that the two rise together and that its data does not show cause27. How an assistant's answers to one question change over days is NF in our data, because our five repeat answers were collected within about 20 minutes on one day4. What a signed-in user sees is also NF, since answers through an API are a baseline2.

What can a local practice take from both?

We think a practice can use the two studies in four ways. Check each assistant by itself, because DataForSEO† reports that platforms favor different sources9 and our assistants seldom named the same business2. Ask the same question more than once, since 13.5 different businesses appeared across five answers4. Look at the listing, review and community sites that both sets of figures show being cited, which DataForSEO† lists as its first priority3. Read click estimates as modeled, as DataForSEO† states they are23.

Our interpretation goes where DataForSEO† does not. We know from the work that an assistant's answer about a local business is largely set by what is published about that business around the web, and that it leans on sites with more authority. We also believe a small business whose listings, reviews and mentions all say the same thing earns that association quickly, while a large one with many competing accounts shifts slowly. That belief is ours, and no figure on this page tests it.

Where does our service fit?

Our plans are $999, $1,999 and $2,999 a month per location, month to month28, and each additional location is $299 a month28. The $999 plan includes a monthly AI answer check, which records what assistants say about a business on the day they are asked. It does not control what ChatGPT, Gemini, Perplexity or Google show, and we promise no ranking or mention.

People also ask

How often do AI Overviews appear in Google results?

DataForSEO† reports that the share of analyzed Google results pages with an AI Overview rose from 5.93% to 61.43% between September 2024 and August 202615. In a separate June to September 2026 panel it reports an overall trigger rate of 56.7%21. It reports Real Estate as the lowest industry at 22.27%13.

Which sites do Google AI Overviews cite most?

DataForSEO† reports that YouTube is the most-cited domain in Google AI Overviews with 51.9M citations, ahead of Reddit, Facebook, Wikipedia and Google, and that Yelp ranks 15th with 2.2M8,20. The snapshot is dated September 14, 2026, for US English searches. Our local questions went to chatbots, where Reddit led in 87 answers2.

Which sites does ChatGPT cite most?

DataForSEO† reports that Wikipedia is the most-cited domain in ChatGPT answers, with 4.88M unique answers, followed by Reddit at 3.25M9. For local questions in our study, the sites ChatGPT cited in the most answers were birdeye.com, in 32 of 199 answers, and bestprosintown.com, in 302.

Do AI answers always cite a source?

No. DataForSEO† reports that 27.3% of ChatGPT answers in its corpus carry no source citation, against 0.7% of Google AI Overview answers6. In our study, with web search on, ChatGPT cited 3.5 websites per answer on average, Gemini 6.5 and Perplexity 7.72.

Do ChatGPT, Gemini and Perplexity recommend the same local businesses?

Seldom. In our study of 199 local questions the three assistants named 2,975 businesses, and only 122 (4.1%) were named by all three, while 79.7% were named by only one2. Each question was asked once through each assistant's API, so the figures are a baseline2.

How much do AI Overviews reduce clicks to websites?

DataForSEO† reports an estimate, built from clickstream data, that AI Overview citations capture 42% to 45% of clicks when an AI Overview is present, and that the combined click share of organic positions 2 to 10 falls by about 36% to 42%15,16. It states that these are modeled estimates and not observed clicks23.

DataForSEO† says Reddit is the second-most-cited domain in both Google AI Overviews and ChatGPT in its datasets, and reports that Reddit appeared in Google's top 10 for 32.15% of searches in Q3 20263,10. In our local-question study, Gemini cited Reddit in 42% of its answers and Perplexity in 0%2.

How to read this page: † a vendor’s own figure, NF not found, ( ) our arithmetic on cited figures, and a small number points to a source below.

Sources

  1. DataForSEO. Search Insider: 2026 Edition, p. 32. DataForSEO, 2025-08 to 2026-09. Secondary source.
  2. Technical Marketing. Who AI recommends locally: a study of local answers. Technical Marketing, 2026-11. technicalmarketing.ai/reports/ai-local-recommendations/ Our own data.
  3. DataForSEO. Search Insider: 2026 Edition, p. 28. DataForSEO, 2026-07-01 to 2026-09-14. Secondary source.
  4. Technical Marketing. Does AI recommend the same businesses twice?. Technical Marketing, 2026-11. technicalmarketing.ai/reports/ai-answer-consistency/ Our own data.
  5. DataForSEO. Search Insider: 2026 Edition, cover. DataForSEO, 2025-07 to 2026-10. Secondary source.
  6. DataForSEO. Search Insider: 2026 Edition, p. 5. DataForSEO, 2025-07 to 2026-09. Secondary source.
  7. DataForSEO. Search Insider: 2026 Edition, p. 31. DataForSEO, 2024-09 to 2026-10. Secondary source.
  8. DataForSEO. Search Insider: 2026 Edition, p. 20. DataForSEO, 2026-09-14. Secondary source.
  9. DataForSEO. Search Insider: 2026 Edition, p. 21. DataForSEO, 2025-07-30 to 2026-09-06. Secondary source.
  10. DataForSEO. Search Insider: 2026 Edition, p. 22. DataForSEO, 2025-10 to 2026-09-14. Secondary source.
  11. DataForSEO. Search Insider: 2026 Edition, p. 29. DataForSEO, 2026. Secondary source.
  12. DataForSEO. Search Insider: 2026 Edition, p. 42. DataForSEO, 2026-06-16 to 2026-09-10. Secondary source.
  13. DataForSEO. Search Insider: 2026 Edition, p. 43. DataForSEO, 2026-06-16 to 2026-09-10. Secondary source.
  14. DataForSEO. Search Insider: 2026 Edition, p. 15. DataForSEO, 2026-06-16 to 2026-09-10. Secondary source.
  15. DataForSEO. Search Insider: 2026 Edition, p. 12. DataForSEO, 2024-09 to 2026-08. Secondary source.
  16. DataForSEO. Search Insider: 2026 Edition, p. 11. DataForSEO, 2026-09-07. Secondary source.
  17. DataForSEO. Search Insider: 2026 Edition, p. 10. DataForSEO, 2025-09 to 2026-08. Secondary source.
  18. DataForSEO. Search Insider: 2026 Edition, p. 25. DataForSEO, 2025-07 to 2026-09. Secondary source.
  19. DataForSEO. Search Insider: 2026 Edition, p. 48. DataForSEO, 2025-07 to 2026-09. Secondary source.
  20. DataForSEO. Search Insider: 2026 Edition, p. 47. DataForSEO, 2026-09-14. Secondary source.
  21. DataForSEO. Search Insider: 2026 Edition, p. 16. DataForSEO, 2026-06-16 to 2026-09-10. Secondary source.
  22. Technical Marketing. What service businesses search for: 2026 demand data. Technical Marketing, 2026-11. technicalmarketing.ai/reports/service-business-search-demand/ Our own data.
  23. DataForSEO. Search Insider: 2026 Edition, p. 37. DataForSEO, 2026-09-07. Secondary source.
  24. DataForSEO. Search Insider: 2026 Edition, p. 27. DataForSEO, 2026-03 to 2026-09-30. Secondary source.
  25. Technical Marketing. South Sound contractor market, 2026. Technical Marketing, 2026-10. technicalmarketing.ai/reports/south-sound-contractor-market/ Our own data.
  26. DataForSEO. Search Insider: 2026 Edition, p. 52. DataForSEO, 2026-03 to 2026-07. Secondary source.
  27. DataForSEO. Search Insider: 2026 Edition, p. 24. DataForSEO, 2025-08 to 2026-09. Secondary source.
  28. Technical Marketing. Pricing. Technical Marketing, 2026-11. technicalmarketing.ai/pricing/ Our own data.

Find out what AI says about you.

A free 20-minute call. We ask the assistants your customers' questions and show you where you come up today.

No pitch unless you ask for one.