Local SEO in the AI Era — Google Business Profile, Reviews and City Pages That Models Recommend

Local SEO is the practice of optimizing a business's visibility for locally-intended queries — "accountant Warsaw", "laptop repair near me" — in Google Maps, the local pack, organic results and, increasingly, AI answers. The stakes are bigger than the AI hype cycle suggests: 46% of all Google searches carry local intent (up from ~30% in 2019), and 76% of people searching for something "nearby" visit a business within 24 hours. Local queries are the shortest path from search to transaction that exists — which is why, despite the AI revolution, the foundations haven't moved: a complete Google Business Profile, reviews and a consistent business entity. What changed is WHO reads those signals: no longer just the maps algorithm, but also the models generating answers and recommendations.

46% of Google searches carry local intent, and 76% of "near me" searchers visit a business within a day. AI changed the interface (AI Overviews on local queries, assistant recommendations), but the foundation stayed: a complete Google Business Profile, reviews and a consistent entity. The complete 2026 local SEO guide — with a step-by-step plan.

This guide walks the whole path: how Google's local algorithm works (from the official documentation), how to complete your profile in 2026 realities (including the services AI now writes in for you), how reviews and city pages work — and how to make ChatGPT and AI Overviews recommend your business.

Why local SEO wins even in the AI era

Two things make local queries special. First, the intent is transactional by nature — whoever searches "plumber Brooklyn" doesn't want an article, they want a phone number. Second, AI can't "consume" this intent with an answer alone: the user still has to visit the business. That's why the click erosion AI Overviews caused in informational content (I cover it in the AI Overviews guide) bypasses much of local search — and the local pack and Maps remain the most valuable real estate in the SERP.

Which doesn't mean AI isn't changing the game: AI Overviews increasingly appear on "best X in [city]" queries, voice and chat assistants recommend specific businesses, and Google began automatically adding AI-generated services to Business Profiles. The signals that built local rankings for years have simultaneously become the training data of recommendations.

How Google's local algorithm works — the three official pillars

Google's official documentation lists three local ranking factors:

1. 1.Relevance — how well the profile and site match the query. This is where profile categories, service descriptions and site content work.
2. 2.Distance — how far the business is from the searcher or the specified location. The only factor you can't influence — beyond having a real presence (and a city page) where you actually operate.
3. 3.Prominence — how well-known the business is across the web: reviews, mentions, links, directory citations. It's the local version of authority — and the strongest lever of advantage.

Industry ranking-factor analyses estimate the weights roughly as: ~32% the Business Profile, ~19% on-page signals, ~16% reviews — the rest is links, behavioral signals and NAP citations.

Google Business Profile — the 2026 setup

A complete profile isn't a formality: businesses with complete profiles get ~70% more location visits than those with incomplete ones. The 2026 checklist:

• Primary category + additional categories — the single most important profile decision; the primary category should match your most valuable query.
• Services — audit what AI writes in. Since 2026 Google auto-generates service lists on profiles based on site content and business data. It sometimes adds services you don't offer — review and correct this section monthly.
• A business description written for the customer, not for keywords — with specifics (specialty, service area, differentiators).
• Photos added regularly (interior, team, work) — profile engagement (photo views, direction requests, clicks) is a ranking signal.
• Posts and updates — a weekly rhythm; a "living" profile beats an abandoned one.
• The Q&A section — answer yourself before a random user does; it's also content the models read.
• Hours, attributes, phone, site link — consistent with the site and directories (NAP consistency below).

Reviews — a ranking signal that became training data

Reviews now work on three levels: as a local pack ranking factor (~16% weight), as a conversion factor (rating and volume decide the click), and — new in the AI era — as content models learn from when deciding whom to recommend. When ChatGPT or AI Overviews recommend "a good dentist in Krakow", they lean on review volume, freshness and text.

The practice that works:

• Systematic acquisition — the review ask built into your service process (after delivery, by email, via QR), not a yearly campaign.
• Review content matters — an opinion mentioning the service and city ("great SEO audit for our Gdansk company") builds relevance more than five stars alone. Don't write reviews for clients — ask questions that naturally elicit specifics.
• Respond to everything — negatives included; responses signal engagement and are content models see.
• Never buy reviews — detection ends in filtered reviews or a suspended profile, and review sellers leave patterns that are easy to detect.

Reviews are the local variant of the mechanism I describe in digital PR and brand mentions: being talked about in credible places is the currency of visibility.

The website and city pages

The profile wins the local pack, but the site decides relevance and conversion. The key elements:

• A dedicated page for every service × location combination where you genuinely operate. I run such city pages myself (process automation per city) — and can confirm first-hand: they work, but only when each carries local value: local case studies, the service area, market context — not the same text with the city name swapped.
• The line between strategy and spam: identical pages differing only by city name are doorway pages — Google filters them. If you have nothing local to say about a city, don't create the page.
• LocalBusiness schema — name, address, phone (NAP), hours, geo, service area, sameAs to your profiles. It's the layer machines read directly; implementation is covered in my Schema.org guide.
• NAP consistency everywhere — identical name, address and phone on the site, profile, directories and social media. Inconsistent data means a blurred entity — and models, as I describe in Entity SEO, must be able to merge every mention into one business profile.

How AI models pick businesses to recommend

When a user asks an assistant to "recommend an accountant for a company in Brooklyn", the model has no business database of its own — it assembles the answer from the same signals that build local rankings: profile and maps data, reviews (volume, ratings, text), presence in local roundups and directories, and site content. The practical consequences:

• Local roundups are the shortcut to AI recommendations — "best accounting firms in Poznan" on a local portal gets cited directly. Auditing your presence in such rankings (and pitching where you're missing) is the cheapest tactic with the fastest payoff.
• A consistent entity decides whether mentions add up — see above: NAP, sameAs, one canonical name.
• Measure recommendations, not just rankings — regularly ask the models "recommend X in [city]" and log who appears; the methodology is in my AI Share of Voice guide.

The step-by-step rollout plan

1. 1.Baseline audit: local pack positions for 10–15 phrases, profile completeness, review count and average, a recommendation test across 3 AI models.
2. 2.Complete the profile: categories, services (correct the AI entries), description, photos, attributes — to 100%.
3. 3.Unify NAP across the site, profile, directories and social media; remove duplicate profiles.
4. 4.Implement LocalBusiness schema on the homepage and city pages (NAP, geo, hours, sameAs).
5. 5.Launch the review process: a standing ask mechanism plus responses to every review within 48h.
6. 6.Build city pages with local value — only for locations where you genuinely operate; each with unique content.
7. 7.A profile publishing rhythm: a post weekly, photos monthly, Q&A continuously.
8. 8.Enter local roundups and directories — industry and city ones with real traffic, not NAP farms.
9. 9.Tie locality into your content cluster: blog answers to local customer questions linking to city pages — the architecture is in my topical authority guide.
10. 10.Measure monthly: the local pack (a local rank tracker), profile insights, reviews, AI recommendations.

The most common mistakes

• An incomplete or abandoned profile. A profile with no photos or posts for a year looks like a closed business — to people and to models.
• Ignoring AI-added services. Google can add services you don't offer — an unaudited section misleads customers.
• Doorway pages. Dozens of identical city pages with a swapped city name are asking for a filter.
• Buying reviews. Short-term gain, long-term suspension risk.
• Inconsistent NAP. Three name variants and two addresses across the web mean a blurred entity and weakened citations.
• No AI measurement. If you don't know whom ChatGPT recommends for your category and city — you don't know who you're really competing with.

Summary

Local SEO is the most underrated channel of the AI era: local intent is growing (46% of searches), converts within a day (76%), and the AI revolution didn't demolish the foundations — it wired them into a new interface. The profile, reviews, consistent NAP and city pages with real value now build your local pack position and your presence in model recommendations at the same time. Two new duties: audit what AI writes into your profile, and measure whom the models recommend in your category.

Strategically: local visibility is entity × trust × local context — the same foundations I describe in E-E-A-T, anchored in geography.

---

I build local visibility for businesses — from profile and entity audits to city pages and AI recommendation measurement — as part of technical SEO and AI optimization (GEO). Get in touch — I'll start by checking whom Google and ChatGPT recommend in your city today instead of you.

Related articles

• Google AI Overviews & AI Mode — optimization
• Digital PR and brand mentions
• Entity SEO and the Knowledge Graph
• Schema.org structured data — the complete guide
• E-E-A-T in 2026 — building trust for Google and AI
• How to measure Share of Voice in AI