Nearly half of all consumers now ask an AI before they ask Google when looking for a local business. That is not a prediction, it is a number from BrightLocal's 2026 Local Consumer Review Survey, which found that 45% of consumers use AI assistants such as ChatGPT, Gemini, or Perplexity to find local recommendations. Travel booking platform data from Adobe Analytics recorded AI referrals to travel and hospitality sites growing 17x year-over-year between 2024 and 2025. SOCi's 2026 Local Intelligence Report found that despite this explosion in AI-driven discovery, only 1.2% of local businesses actually appear in AI-generated responses to local queries. The other 98.8% are invisible, not penalised, not ranked low, simply absent. This article breaks down exactly what is driving the shift, which businesses are winning, and the practical location-data steps that move you from invisible to cited.
[Image: Bar chart comparing consumer use of AI for local recommendations: 2023 (~8%), 2024 (~22%), 2025 (~35%), 2026 (45%), sourced from BrightLocal data]
The Numbers Behind the Shift
The statistics are moving fast enough that figures from 18 months ago already feel historic. Here is the current picture:
- 45% of consumers use AI for local recommendations (BrightLocal, 2026)
- 17x growth in AI referral traffic to travel sites in a single year (Adobe Analytics, 2025)
- 1.2% of local businesses appear in AI local query responses (SOCi, 2026)
- 62% of AI-assisted local searches do not result in a follow-up Google search, the consumer acts directly on the AI's recommendation
- 3.4x higher conversion rate from AI referral traffic compared with organic search traffic (Adobe Analytics, 2025)
The last two figures matter most for revenue. When an AI recommends your business, the person asking has already narrowed their intent to a single query. They are not browsing, they are deciding. The click that follows is worth more than a typical organic visit, and it never appears in your Google Search Console data.
Why 98.8% of Businesses Are Invisible to AI
The gap between 45% consumer adoption and 1.2% business representation is not an algorithm penalty. There is no list AI models consult to decide who to exclude. The absence happens because AI models require high-confidence structured signals to cite a specific business, and most businesses have never provided them.
Missing Structured Data
AI models parse the web continuously. When they encounter a business website that contains only prose, "We're a family-run Italian restaurant in Lyon serving seasonal dishes since 1998", they extract fragments. When they encounter a website with a properly implemented LocalBusiness JSON-LD block containing the business name, address, latitude/longitude, opening hours, and price range in a machine-readable format, they can resolve the entity with confidence. The difference between being cited and being ignored often comes down to a single <script> tag in the HTML <head>.
To learn which fields matter most for AI citations, see our guide to JSON-LD schema for local businesses.
NAP Inconsistency
Name, Address, and Phone number must match exactly across every source an AI model can reach: your website, your Google Business Profile, TripAdvisor, Yelp, Facebook, and relevant local directories. A business listed as "Café du Marché" on its website but "Cafe du Marche" on Yelp and "Café Du Marché SARL" on its Google Business Profile is three different entities from an AI model's perspective. None of them accumulates enough corroborating signal to cross the confidence threshold for citation. We cover this in detail in NAP consistency for AI search.
Review Freshness and Volume
AI models weight recency. A business with 200 reviews, the most recent from 14 months ago, is less citable than a business with 40 reviews, the most recent from last week. The model interprets recent reviews as a signal that the business is actively operating and that its information is current.
The Industries Where the Shift Is Happening Fastest
The 45% headline is an average. In some categories, AI adoption for local discovery is already the majority behaviour:
- Restaurants and cafés: 58% of consumers aged 18–34 used AI to find a restaurant in the past 90 days
- Hotels and accommodation: AI travel query volume grew 340% in 2025; 80% of travellers now use AI at some stage of trip planning
- Healthcare providers: 41% of patients used AI to find a GP, dentist, or specialist in 2025
- Home services: plumbers, electricians, and cleaners are the fastest-growing AI local search category
The businesses winning in these categories are not necessarily the largest or the best-reviewed. They are the ones whose structured data is complete enough for AI models to confidently recommend them.
What AI Engines Actually Look For
Understanding what these models need makes the fix feel less abstract. When a user asks ChatGPT "best Italian restaurant near me open on Sunday evenings in Porto," the model is not running a live search in the way Google does. It is pattern-matching against a large corpus of structured knowledge. The businesses that appear are those whose data was unambiguous, consistent, and well-structured when that corpus was last updated.
The key signals are:
- Precise geocoordinates, latitude and longitude in schema markup allow the model to resolve "near me" queries accurately
- Opening hours in structured format,
openingHoursSpecificationin JSON-LD, not just prose text - Service area or geographic coverage, especially for businesses that serve multiple neighbourhoods or cities
- Category and cuisine/specialty markup,
@type,servesCuisine,priceRange - Consistent cross-web presence, the same entity appearing in authoritative directories with matching information
This is exactly the signal pipeline described in our complete guide to AEO (Answer Engine Optimization).
[Image: Screenshot of a ChatGPT response to "best Thai restaurant open Sunday in Amsterdam" showing three businesses with name, address, opening hours, and price range, demonstrating what a fully cited business looks like vs. a vague mention]
The Revenue Connection
The conversion data is the reason to care about this beyond vanity metrics. Adobe Analytics found that visitors arriving via AI referral convert at 3.4x the rate of organic search visitors. This is intuitive once you consider the query context: someone who asked an AI for a specific type of business in a specific area and received your business as the recommendation has already completed most of their decision-making process. They are not in the discovery phase, they are in the commitment phase.
For a restaurant with 20 covers per service, moving from invisible to cited in AI responses for even a modest number of daily queries translates directly to more reservations. For a hotel, the same shift affects room-night bookings. The economics of AI visibility are not subtle.
Four Practical Steps to Take This Week
The gap between the 1.2% who appear and the 98.8% who do not is a solvable technical problem, not a years-long campaign.
Step 1: Audit your current AI visibility. Use the free MapAtlas AEO Checker to scan your website's structured data, NAP consistency, and location signals in under 60 seconds.
Step 2: Implement or fix your JSON-LD schema. Add a complete LocalBusiness block to your site's <head>. Include geo (coordinates), openingHoursSpecification, priceRange, servesCuisine (if applicable), and sameAs links to your authoritative profiles. Full markup example in our JSON-LD schema guide.
Step 3: Audit NAP consistency. Check your business name, address, and phone number across your website, Google Business Profile, Apple Maps, TripAdvisor, Yelp, and Facebook. Fix any discrepancies, even minor formatting differences.
Step 4: Publish location-specific content. A 400-word page describing your neighbourhood, nearby landmarks, parking, and what makes your location distinctive gives AI models context they cannot get from schema alone. Update it when hours or services change.
The Window Is Still Open
The 45% figure will keep rising. Consumer habits around AI-assisted local search are following the same adoption curve as mobile search did a decade ago, and businesses that moved early on mobile search captured audience that competitors never recovered. The structural advantage of being in the 1.2% now is that you establish citation precedent in AI training data while your competitors are still deciding whether to act.
The MapAtlas AI Search Visibility solution is built specifically for this transition, connecting the structured geodata signals that AI engines require with the monitoring and verification tools businesses need to maintain them. If you are ready to move from invisible to cited, start with a free audit today.