Why ChatGPT Doesn't Recommend Your Local Business (and the 30-Minute Saturday Fix)

Customers ask ChatGPT for a plumber and get three names. Yours probably isn't one. The pipeline is Foursquare + Bing — not Google. The 30-minute Saturday fix, sourced.

It’s a Tuesday at 8 PM. A homeowner in your city has a leaking water heater. Three years ago, they would have typed “emergency plumber [your city]” into Google. Tonight, they open ChatGPT and ask:

“Best emergency plumber in [your city], doesn’t gouge on weekend rates?”

ChatGPT thinks for a second. It returns three businesses. Maybe four.

That’s the new front page of search.

For most local businesses, when ChatGPT names three plumbers in their city, they aren’t on the list. Not because the business is bad. Because nobody told the owner how the new game works.

This piece is the 30-minute version of how the new game works. It’s the same fix our AI Visibility for Local Businesses course walks through end-to-end, condensed to what you can run before lunch on a Saturday.

The Number That Should Get Your Attention

ChatGPT is now handling around 2.5 billion queries per day globally — about 330 million from US users alone, per OpenAI’s own July 2025 disclosure. Perplexity is in the tens of millions of weekly queries, and Google’s AI Overviews now appear in roughly 48% of searches as of early 2026 (up from 34.5% just four months earlier, per Indexly’s analysis).

For high-intent queries — who do I call right now? — conversational AI is rapidly taking share. And critically: customers don’t see ten blue links anymore. They see three names.

If you’re not on the three-name list, you’re not in the conversation.

The Reveal: ChatGPT Isn’t Pulling From Google

This is the part most local business owners — and most “AEO agencies” charging $1,500–$8,000 per month — don’t say out loud.

When ChatGPT serves a local recommendation, the data flows like this:

  1. Foursquare’s Places API returns the canonical business profile (name, category, address, hours, photos, ratings)
  2. Reprompt, an LLM-based agent that Foursquare officially announced as their “Placemaker Agent” in January 2025, has already enriched that profile with fresher data scraped from the business’s website and social profiles
  3. Bing Search layers real-time signals on top — recent reviews, today’s hours, news mentions

Industry analyses (GMBapi, Superprompt) estimate Foursquare provides 70%+ of the data signal for local recommendations on ChatGPT. The remaining ~30% is Bing real-time + the enrichment layer.

Notice what’s missing from that pipeline. Google. Your Google rank means basically nothing here.

A separate analysis (ekamoira, summarizing Status Labs research) found that only 12% of URLs cited by ChatGPT also rank in Google’s top 10 organic results. You can be #1 on Google for “plumber [your city]” and ChatGPT will quietly recommend your competitor. Different data sources, different rankings.

How Perplexity and Gemini Differ

Perplexity is web-first. It crawls the open internet and cites high-trust sources inline. For local queries, that means:

  • Reviews aggregators like Yelp, TripAdvisor, MapQuest, and BBB
  • Reddit threads — Reddit accounts for ~24% of Perplexity citations in some categories per the Pikaseo report (February 2026), and Reddit’s citation share grew 73% between October 2025 and January 2026
  • First-party websites with strong on-page entity signals (clear NAP, schema markup, service descriptions)

A Yext analysis of 6.8 million AI citations (summarized in Search Engine Land, October 2025) found 86% of all AI citations come from sources brands control — 44% from first-party websites and 42% from listings. Only ~2% come from forums other than Reddit.

Gemini is the simplest of the three: it’s the most integrated with what you already know. Gemini lives inside Google Maps, Google Business Profile, and the same organic web index that powers Google Search. If your GBP is tight, you’re probably already showing up in Gemini.

The uncomfortable truth: the three big AI engines pull from largely non-overlapping data sources for local queries. Optimizing for one doesn’t automatically lift the others.

The 30-Minute Saturday Audit

Here’s the rapid version. Each block is short enough to do alone, between coffee and lunch.

1. Mirror-test your business (5 min)

Open three tabs: chat.openai.com, perplexity.ai, gemini.google.com. In each, type:

“What are the 3 best [your trade] in [your city] for [common high-intent need]? Give names, what they’re known for, and how to contact them.”

Screenshot every answer. Save them in a folder named “AEO baseline – [today’s date].” This is your before-state. You’ll re-run the same prompts in 4 weeks and 12 weeks to measure progress.

If your business doesn’t appear in any of the three engines: you’re invisible to AI search right now. That’s the most common starting point. The next four blocks fix it.

2. Claim your Foursquare profile (10 min)

This is the step nobody sells you on. Go to foursquare.com/explore, search your business name + city. Three things might happen:

  • The listing exists and is partly accurate → claim it via business.foursquare.com
  • The listing is unclaimed → claim it
  • No listing exists → create one

Verification is by phone or postcard (postcard takes 5–10 days; that’s fine). Then fill out every field: NAP, primary category (most specific available — “Pediatric Dentist” not “Dentist”), hours with weekend variations, services with descriptions, price range, photos (8–12 real ones, not stock).

Profile completeness is the single biggest driver of ChatGPT’s ranking confidence. Sparse profiles lose to complete ones, every time.

3. Tune up Google Business Profile (10 min)

GBP drives Gemini and Google AI Overviews. The fields most SMBs underuse:

  • Services list — each with name, 1–3 sentence description, price when relevant. AI Overviews cite these directly
  • Business description — 600–700 of your 750 character allowance, plain language, no hype words
  • Q&A — pre-empt 5–8 common customer questions and answer them yourself; AI Overviews pull this content directly
  • Posts — at least monthly, ideally weekly. Mini-announcements are read as activity signals

4. Bing Places (5 min — really)

Customers don’t use Bing. ChatGPT does. Microsoft’s Bing powers ChatGPT’s real-time web layer, and Bing Places is the canonical Bing index entry for your business.

Go to www.bingplaces.com. Click “Import from Google.” Bing pulls your GBP data and creates a Bing Places listing in about ten minutes. Verify ownership, fix any drift Bing imported imperfectly, done.

This is the highest-ROI five-minute task in the audit.

5. NAP consistency sweep (5 min — once you’ve done the above)

Walk through your major listings — website footer, Google Business Profile, Bing Places, Foursquare, Yelp, Apple Maps, Facebook, BBB — and confirm Name, Address, Phone are character-identical everywhere. “123 Main Street” and “123 Main St” are different signals. Pick one canonical format and propagate it.

That’s the 30-minute foundation. The Saturday-morning version of our AI Visibility for Local Businesses course adds LocalBusiness JSON-LD schema (a copy-paste block of code for your homepage) and a comparison guide — the highest-leverage content move.

The Single Highest-Leverage Content Move

After the 30-minute fix, the most impactful next step isn’t more directory submissions or more schema. It’s one comparison guide for your local market.

The Digital Bloom 2025 AI Citation Visibility Report (April 2026), based on 680 million AI citations, found that comparative listicles account for 32.5% of all AI citations — the single highest content format share.

A working comparison guide for your city looks like “2026 [Your City] [Your Trade] Cost & Service Guide” and includes:

  • A neutral overview of how to choose a [trade] in your city
  • A real cost-breakdown table for the most common services
  • A comparison of 5–7 actual local competitors (your business included, equally and neutrally)
  • An honest FAQ
  • A “when to call vs DIY” section
  • A monthly “last updated” stamp

Multiple operators have shared real before-and-after results from this exact pattern:

  • A plumber published a “Local Plumbing Cost & Service Guide” on Medium with a $200 press release. Within weeks, ChatGPT was recommending him for emergency plumbing queries in his city. (@boringlocalseo, March 2026)
  • A real estate agent published “2026 Nashville Real Estate Market Report: Neighborhood Analysis” with a top-rated agents comparison. ChatGPT and Perplexity recommendations followed within a month.
  • A plumber sent 3 press releases over 2 months. Result: AI recommendations + 2 booked jobs from AI referrals. (@noahiglerSEO, February 2026)

The pattern is consistent. Comparison guide + Medium + light PR distribution → AI citations within 2–6 weeks. No agency. No five-figure retainer.

The Myth Roll Call

While we’re here: a few tactics being actively sold by AEO agencies right now that don’t work. Don’t pay for these.

llms.txt files. Tested across 40 sites — only 12% saw any citation increase after 60 days (@geo_profesor, March 2026). Google executives have publicly dismissed it. No major AI company endorses it.

“AI directory submissions.” A category invented by the people selling the service. AI engines don’t crawl “AI business directories” with priority over normal directories. Most are zero-traffic single-page-rank sites.

FAQ schema on every page. One SE Ranking study (cited in late-2025 Stackmatix research) found pages with FAQ schema averaged 3.6 ChatGPT citations vs 4.2 without — a small negative when the schema was added cosmetically without real FAQ content.

Keyword density tactics. A 2026 analysis of 1,548 brands (@geobuddyco, March 2026) found zero correlation between AI visibility and keyword density, llms.txt, or schema markup added without context. What did correlate: review presence, Wikipedia mentions, and genuine category authority.

If your AEO agency proposes any of the above as their differentiator, ask them to show you the citation data. They can’t, because it doesn’t exist.

What Honest Pricing Looks Like

Industry vendor data (TryAivo, Stackmatix, late 2025–early 2026):

Company sizeTypical AEO monthly spend
Budget SMB / single location$1,500–$3,000
Multi-location SMB$3,000–$8,000
Mid-market$8,000–$25,000
Enterprise$25,000+

For a single-location SMB, the entire foundational audit (Foursquare + GBP + Bing + schema + comparison guide) is something you can run yourself in a Saturday morning. If you hire someone, the right ask is “we built our own foundation; help us with scaled PR distribution and entity-graph cleanup” — that’s $1.5–$3K/mo of real value, not $8K of repackaged Google SEO.

What This Is NOT

A few honest caveats, because you’ve been pitched on this before:

  • No “guaranteed top placement.” AI rankings shift weekly. Anyone selling a guarantee is selling fiction.
  • No instant results. Citations take 2–6 weeks to appear; visible referrals take longer.
  • No substitute for measurement. Tag your own funnel with UTMs, set up call tracking if you can, ask new customers “how did you find us?” That’s the only data that matters for your business.

There also isn’t yet rigorous public data for several stats that float around AEO marketing copy — “X% of US consumers use ChatGPT for near-me queries,” “AI traffic converts at 24% on Webflow,” “the 0.334 brand search correlation.” When we built the AI Visibility for Local Businesses course, we explicitly flagged each of those as unverifiable to original methodology. The honest message: nobody has perfect numbers yet. You and your competitors are running the same experiment in real time.

What Comes After Saturday

If you do nothing else: claim Foursquare. That’s the 10-minute action with the largest single payoff. Most SMB owners we’ve audited had never logged into Foursquare. After they do, ChatGPT visibility shifts within 2–4 weeks.

If you want the full Saturday-morning audit — including the JSON-LD schema you can copy-paste into your homepage, the comparison-guide outline and AI prompt that produces a strong first draft, the review request system that adds 5 reviews a month, and the measurement plan that tells you what’s actually working — that’s what the AI Visibility for Local Businesses course covers across 8 lessons.

For broader local SEO foundations alongside AEO, SEO Mastery is the natural pair. For wider AI workflows beyond visibility, AI for Small Business covers operations, hiring, customer service, and more.

But the core message: stop paying for llms.txt setups. Start with Foursquare. Ten minutes. This Saturday.

Your customers are already asking ChatGPT. The question is whether your business is one of the three names that comes back.

Build Real AI Skills

Step-by-step courses with quizzes and certificates for your resume