Local AI Search by the Numbers: 4 Years of Data on the Shift

Nearly four years after ChatGPT launched, the question is no longer whether AI search matters for local businesses in your market. The data settled that argument some time in 2024. The interesting question now is how much, where, and what to do about it. This post pulls together the verified third-party data from the last four years into one place, with each statistic tied to a named study.

Adoption has been steep, and it is still climbing

Pew Research's February 2026 survey of 5,119 US adults found that 44% of US adults have used ChatGPT, up from 34% a year earlier and more than double the share in 2023, and 49% have used at least one AI chatbot. Adults under 50 are about twice as likely as those 50 and older to have used ChatGPT (57% vs 28%), and 38% of employed adults use chatbots for tasks at work.¹

OpenAI's own weekly-active-user figures are higher than any consumer-product adoption curve a decade ago would have predicted. ChatGPT crossed 200 million weekly active users in August 2024; 300 million by December 2024; 400 million in February 2025; around 700 million in early August 2025;² 800 million in October 2025;³ 900 million in February 2026;⁴ and, as reported in August 2026, more than 1 billion, per OpenAI's own announcements.⁵

For comparison, ChatGPT.com's monthly web visits more than doubled, from 2.6 billion to 5.7 billion in the twelve months to July 2025, then levelled off into 2026, per Similarweb.⁶ Perplexity, a smaller but widely cited AI search product, grew from 10 million MAU in January 2024 to 30 million by April 2025 and roughly 45 million by the end of 2025.⁷ Google's Gemini app is far larger: Google said it passed 1 billion monthly active users in August 2026.⁸

In one local study, AI Overviews appeared on 2 in 3 local-business queries

Google's AI Overviews launched in the US in May 2024 and have since rolled out across most major markets. Tracking the prevalence of AI Overviews has become a small industry in itself. Two data sets are most useful here:

• Semrush analysed over 10 million keywords through 2025 and found AI Overviews peaked at 24.61% of queries in July 2025, then settled at around 16% (15.69% in November 2025).¹⁰
• Whitespark's local-business study is the cleanest cut for our purposes. Across 540 queries in six local-business verticals in three US cities (published May 2025), AI Overviews appeared on 68% of local-business queries on average, but the rate varies enormously with intent shape.¹¹

What happens to clicks when an AI Overview appears

Several studies from 2025 and early 2026 measured the click-through-rate impact of AI Overviews.^12131415 Each used different methodology; each came back with a meaningful drop.

The directional finding is unambiguous: when an AI Overview appears, the top organic result loses meaningful click share. The size of the loss depends on query type, sector, and how the AIO is rendered. None of the major studies have published local-search-specific CTR data yet, which is the most obvious gap in the public evidence base.

SparkToro's zero-click research puts the broader context in numbers. Its 2024 study (Datos clickstream data) found 58.5% of US Google searches ended without any click, and only about 360 of every 1,000 US searches sent a click to the open web.¹⁶ Its June 2026 update, using Similarweb clickstream data for January to April 2026, found that 68% of US Google searches now end without a click. AI Overviews have accelerated a trend that was already well underway.¹⁷

Ranking number one on Google in 2026 is not what it was in 2022. The visit-per-rank rate has been falling for years; AI Overviews have made the slope steeper. The defensible position is being one of the cited or recommended sources inside the AI answer itself, not just the top blue link below it.

Nearly half of consumers now ask AI for local recommendations

Survey data from 2026 puts the consumer side of the shift in concrete numbers. BrightLocal's Local Consumer Review Survey 2026 (1,002 US adults) found:

ChatGPT is not merely the most-used AI assistant; it is the most-used AI assistant for local-business discovery, ahead of Google's AI Mode.¹⁸ An adjacent survey for Omnisend (1,224 US consumers, July 2025) found that 59% of US shoppers use generative AI tools for shopping tasks, and 1 in 4 say ChatGPT's product recommendations are better than Google's.¹⁹

In BrightLocal's 2026 survey, millennials were the most likely to use AI for local recommendations, and adults over 60 were far less likely.¹⁸ The split matters for sector planning: a hospitality or beauty business serving a 25-45 demographic is seeing AI-discovery behaviour today; a will-writing or hearing-aid practice serving a 65+ demographic is on a longer adoption tail.

AI referral traffic now converts better than other traffic

Adobe Analytics has been publishing quarterly snapshots of generative-AI referral traffic to US retail sites since late 2024. The growth numbers are extreme; the engagement numbers are more interesting.

The same Adobe data set tracks raw growth: AI referral traffic to US retail sites was up 1,300% year-over-year in November-December 2024,²¹ 4,700% year-over-year in July 2025²² and 693% over the 2025 holiday season.²³ Growth has slowed as the base has grown: up 62% year-over-year in July 2026.²⁴

The engagement-quality picture is the more useful one for budget conversations:

The four-year timeline that got us here

For anyone trying to make sense of how fast this all moved, here is the milestone sequence with the actual launch dates. Most of the arrival happened in a 24-month window.

1. 1

   November 2022: ChatGPT launches

   OpenAI releases ChatGPT to the public on 30 November 2022. It crosses 100 million users in two months, the fastest consumer-product adoption ramp on record at the time.

2. 2

   February 2023: Bing Chat launches

   Microsoft launches its GPT-4-powered chat experience inside Bing. It is renamed Copilot in November 2023, and its answers are grounded in Bing's index.

3. 3

   May 2024: Google AI Overviews launches in the US

   Google rolls AI Overviews out of Search Generative Experience (SGE) and into the live US SERP. Early answers draw criticism and Google adjusts when they appear. In Semrush's tracking, coverage then grows through 2025 to a July peak before settling at around 16% of queries.

4. 4

   June 2024: Apple Intelligence announced

   Apple announces Apple Intelligence at WWDC 2024 with ChatGPT integration on the roadmap. The implication for local SEO: your Apple Business listing (formerly Apple Business Connect) becomes worth the same effort as Google Business Profile for iOS-heavy audiences in any market.

5. 5

   December 2024: ChatGPT integrates with Siri

   iOS 18.2 ships on 11 December 2024 with the Siri-to-ChatGPT handoff. Voice queries that Siri cannot handle natively can now route to ChatGPT for answer.²⁵

6. 6

   March 2025: Google AI Mode rolls out

   Google launches AI Mode as a more conversational, fan-out-driven alternative to AI Overviews. Initially limited to Labs; widely available later in 2025. AI Mode is the surface where query fan-out fundamentally changes which pages get cited. See the AI search visibility guide for the mechanism.

7. 7

   July 2025: AI Overviews coverage peaks

   Semrush's 10-million-keyword tracker measures AI Overviews on 24.61% of queries, the high-water mark before coverage settles at around 16% over the following months.

8. 8

   2026: the assistants reach a billion

   OpenAI reports 900 million weekly active users in February 2026, more than double the 400 million figure from a year earlier, and is reported to have passed 1 billion by August. Google's own AI surfaces operate at a similar scale: at I/O in May 2026 Google said AI Mode had passed 1 billion monthly users and AI Overviews reach 2.5 billion, and in August the Gemini app passed 1 billion monthly users.

What the data implies for local SEO investment

Looking across the studies above, six concrete budget priorities emerge. Each one is directly supported by one of the data points cited earlier in the post.

The order above is roughly empirical: items higher in the stack have stronger public-data support for impact and lower-cost implementation. Notably absent from this list: paid placements (ChatGPT now shows labelled sponsored ads below some answers in the US and UK, but they sit outside the answer itself), "AI-optimised content" stuffed with AI jargon, and any approach that treats AI search as a separate channel from traditional SEO. The data does not support any of those framings.

What the data does NOT tell us yet

A complete read of the evidence has to flag the gaps. Six honest limitations to keep in mind when reading any of the numbers above:

Where this leaves us in 2026

The honest summary: AI search is now large enough to matter, distributed across enough surfaces to be hard to game, and growing fast enough that any plan written before late 2024 needs a rewrite. The data does not support panic (Google is not dying), but it does support reallocation. A local SEO budget that still spends 80% on Map Pack levers and 20% on everything else is allocating against the surface distribution of 2021, not 2026.

For a practitioner-level walk-through of what to actually do (entity convergence, page-level schema, the GEO levers in priority order, sector heuristics, and the 90-day programme), the place to go next is the long-form AI search visibility guide. This post quantifies the shift; that guide is the playbook for responding to it.

Where to go next