Local GEO: Getting Recommended for 'Near Me' Prompts

#01How do assistants handle "near me"?

An assistant needs to know two things: where the person is, and which businesses near them fit the request.

Location. Google explains that it can estimate your location from your device's location, your IP address and your past activity (Google). OpenAI says ChatGPT search can use approximate location derived from your IP address to make results more relevant (OpenAI). So the same prompt typed in two neighbourhoods can get two different answers. That matters for testing, which we cover below.

Businesses. Google describes AI Mode as drawing on high-quality web content plus "fresh, real-time sources like the Knowledge Graph, info about the real world, and shopping data for billions of products", using a query fan-out technique across subtopics and data sources (Google). For local questions, "info about the real world" plainly includes the business data behind Maps and Business Profiles. Assistants outside Google lean more on web search results, directories, review sites and your own pages.

#02What signals matter, and which can you control?

Google's documentation for local results is the clearest public statement of what matters. Local results are mainly based on relevance, distance and prominence (Google).

Our view: assistants add a fourth practical factor: answerability. A prompt like "dentist near me who does same-day crowns and takes my insurance" can only be answered for businesses whose crowns and insurance policies are written down somewhere crawlable. Most local sites don't say either.

#03What's the local evidence checklist?

On Google Business Profile

• Verified profile, with the business name as used in the real world (Google)
• Primary and secondary categories that match what you actually do
• Hours, including special and holiday hours (Google)
• Attributes such as accessibility, parking and payment options
• Services or products with descriptions
• Photos that show the real premises, team and work
• Replies to reviews

See Google Business Profile and AI Mode for the profile in detail.

On your website

• One page per core service, with the service name in the heading and the town or area it covers
• Prices or price ranges, and what affects them
• Specialties, brands or methods you use, and who you're not right for
• Booking, emergency and out-of-hours details
• Consistent name, address and phone (NAP) matching the profile
• LocalBusiness or Organization structured data where accurate (Google)

Off your website

• Reviews on Google and on the platforms your category uses
• Local news, community and partner mentions
• Relevant directories and associations, kept consistent (see directory listings)

Ask for reviews honestly. Google's Maps content policy prohibits fake engagement, including paid reviews and incentives for posting reviews (Google).

#04Which prompts should a local business track?

Write prompts the way people speak, and cover different intents:

Fifteen to thirty prompts is enough for a single-location business. Multi-location businesses need a set per location. See building a prompt set.

#05How do you test "near me" fairly?

Location is the biggest source of noise in local testing.

1. Fix the test location. Test from the same place, or set a consistent location, every time, and record it (Google).
2. Test where customers are. A business on the edge of town may look invisible from its own premises' viewpoint if most customers search from the centre. Pick two or three representative points.
3. Repeat runs. Run each prompt several times; answers vary between runs. See why AI answers differ.
4. Log the sources. Note which directories, review sites and articles assistants cite. Those are where your effort should go.
5. Separate signed-in and signed-out results. Personal history and memory can shape answers (OpenAI).

#06Which local pitfalls should you watch for?

• Keyword-stuffing the business name. Google's guidelines say the name should reflect the real-world business name (Google).
• Service-area confusion. Listing a street address for a business that serves customers at their location, or the reverse, against Google's guidelines.
• Stale hours. The most common wrong fact in local answers, in our experience, and the cheapest to fix.
• Thin location pages. Dozens of near-identical "plumber in [town]" pages are a pattern Google's spam policies warn against (Google). Write fewer, genuinely local pages.

#07What's a realistic timeline?

Profile fixes can be reflected quickly. Reviews and local mentions take months. The 90-day local GEO plan template breaks the work into phases and shows how to re-test.

#08From MentionX data: how often brands appear by prompt typeOriginal data

Method. Read-only aggregate queries over MentionX scans from 22 August to 29 October 2026 (UTC): 2,217 engine runs (2,154 completed) across 18 anonymised projects and six engines.

The MentionX sample is mostly B2B software, not local businesses, so it can't show local results directly. It does show how much the prompt type changes whether a brand is named:

"Best plumber near me" is a best-of or recommendation prompt, and in the sample those mention any brand far less often than brand-specific prompts. Engines also differ in how often they cite sources at all:

Our view: local businesses should expect wide variation between assistants and between runs, and should track both discovery prompts and prompts that name the business.

Limits. A convenience sample of mostly B2B software projects, run through engine APIs, which can behave differently from the consumer apps. Perplexity, Copilot and DeepSeek figures come from a single day (26 October 2026), so treat them as a snapshot, not a trend.

#09Worked example: a fictional plumbing firm that uses Ledgerline

Ledgerline is a fictional accounts-payable tool. Figures are illustrative.

Ledgerline's customer Brightwater Plumbing (fictional) asks why it never appears for "emergency plumber open now near me". Its Business Profile lists weekday hours only, its website has a single "Services" page and it has few recent reviews.

Brightwater adds out-of-hours details and holiday hours to the profile, picks categories that match its real work and replies to every review. On its site, it creates separate pages for emergency repairs, boiler servicing and heat pump installation, each with the area covered, price ranges and booking details. It asks recent customers for honest reviews, with no incentive.

It tests ten prompts from two fixed points in town, three runs each, and logs which directories and review sites assistants cite. Two of those directories had an old phone number, which it corrects.

#10How should multi-location businesses organise this?

Give each location its own Business Profile and its own page on your site, with the address, hours, phone number, services and staff for that branch. Avoid copying the same text across location pages with only the town name changed. Include something genuinely local on each page, such as parking, nearby landmarks or the services that branch offers. Keep a single sheet of every location's facts so updates reach the profile, the site and directories at the same time.

#11How do you write a service page assistants can use?

Start with one sentence saying what the service is, where you offer it and who it's for. Add price ranges or what affects the price, how quickly you can attend and how to book. List what you don't do, so assistants don't recommend you for the wrong jobs. Answer the follow-up questions customers usually ask by phone. Keep the name, address and phone number on the page consistent with your profile.

#12How do reviews help beyond the star rating?

Review text often describes the specific job, the area and the speed of service, which is the detail that matches specific prompts. Our view: asking customers to mention what you did for them, without scripting or rewarding them, produces more useful reviews than asking for five stars. Reply to reviews with facts, and fix the problems they raise.

#13What should you record in a local test log?

Record the date, the prompt, the assistant, the test location, whether you were signed in, whether your business was named and which sources were cited. Note any wrong facts, such as old hours. Over time, the log shows which fixes changed answers and which directories assistants rely on most for your area.

#14How do you keep facts consistent across directories?

List every directory and platform that shows your business, including maps apps, review sites and industry associations. Check the name, address, phone number, hours and website link on each. Update them from a single facts sheet, and record when each was last checked. Our view: assistants often cite directories for local prompts, so a wrong phone number on one can end up in an answer.

#15What should you do when an assistant gets your facts wrong?

Find the source the assistant cites, if it shows one, and correct the fact there. Update your profile and website so the correct fact is clear. Re-test after a few weeks. If the error persists and the assistant offers a feedback option, report it with the correct fact and a link. Keep the log in a shared sheet so everyone working on the business can see what changed and when.

#16Common questions

Do 'near me' results differ by assistant? Yes. Each assistant uses its own sources and location estimate, so test each one you care about separately.

Should we create a page for every nearby town? Our view: only where you genuinely serve that area and can say something specific about it. Near-identical town pages risk breaking Google's spam policies.

How often should a local business re-test? Our view: monthly, and after any profile or website change.