How to run a local GEO baseline audit

Ask ten local business owners how they are doing in AI search, and nine will point to their Google Business Profile. That is the wrong place to look. According to SOCi's 2026 Local Visibility Index, ChatGPT recommended only 1.2 percent of nearly 350,000 analyzed locations. Those same brands appear in Google's local 3-pack 35.9 percent of the time — a gap of roughly thirtyfold. Gemini recommended 11 percent of locations, Perplexity 7.4 percent. Business profile information was only about 68 percent accurate on ChatGPT and Perplexity, versus 100 percent on Gemini, which relies entirely on Google Maps data.

A business can dominate the map pack and still disappear the moment someone asks ChatGPT for a recommendation. Many local businesses have never checked what AI says about them. They invest in content and citations without knowing whether any of it shows up where visibility now matters. A local GEO baseline audit closes that gap: it provides a repeatable way to measure how AI describes, recommends, or ignores a business before more budget is spent.

Why the baseline must come first

Without a starting number, progress cannot be proven. The baseline delivers metrics for share of voice, citation rate, and factual accuracy. It also answers a strategic question: Can AI even crawl, understand, and trust this site? The answer shapes everything that follows. Eligibility issues must surface before any content strategy.

AI weights signals differently from classic local search. Traditional local SEO leans heavily on proximity. AI prioritizes data confidence, authority, and consistency across the web. Third-party validation and consistent business data often matter more than how far away the searcher is. That is why map-pack rankings tell you almost nothing about AI visibility.

Step 1: Assemble your audit inputs

Before running a single prompt, structure a spreadsheet. Four query categories expose different weaknesses:

• Discovery: "best [service] near me" or "top [service] in [city]"
• Comparison: "[Brand] vs. [Competitor] in [city]"
• Trust: "[Brand] reviews" or "Is [Brand] reliable?"
• Logistics: hours, address, parking, and phone number

Run each query across the platforms customers actually use: ChatGPT, Perplexity, Gemini, and Google AI Overviews. Each platform pulls from different sources and phrases answers differently. Appearing on one surface does not guarantee appearing on another.

Control for noise: test from a defined location and note the exact city or ZIP code. Run a clean, logged-out session alongside a logged-in one to reduce personalization. Date-stamp every run. Models update constantly; a screenshot without a date is barely comparable.

Step 2: Run the prompts and record the results

For every prompt on every platform, capture five points:

• Mention: Did AI mention the business by name?
• Mention order: First, middle, last, or missing?
• Sentiment and framing: Positive, neutral, or negative?
• Factual accuracy: Are hours, services, and prices correct?
• Cited sources: Which URLs and directories did the answer rely on?

Recommended spreadsheet columns are prompt, platform, mention, position, accuracy score, sentiment, citation count, and top sources. From these, calculate two summary metrics: visibility percentage (how often the business appears) and accuracy percentage (how often the facts are correct). Those numbers form the baseline for later comparisons.

A share-of-voice comparison also helps: how often are direct competitors named in the same discovery and comparison prompts? That shows whether the brand is missing, appears only rarely, or systematically trails more strongly cited alternatives. Save screenshots and raw answers with timestamps so later model updates stay traceable.

Step 3: Check eligibility and data quality

If AI rarely mentions the business, the issue often sits before content production. Check crawlability for relevant bots, consistent NAP data across directories, and whether structured data makes the business type unambiguous. Contradictions in address, phone, or hours lower data confidence. Gemini may be correct via Maps data while ChatGPT and Perplexity mirror outdated third-party sources — that difference is exactly why the baseline audit matters.

Competitor and source picture

Note which competitors are recommended in the same prompts and which domains serve as sources. Industry portals, review platforms, and local media often outweigh the brand's own site. That shows where authority signals are missing and where citations or PR must catch up before further on-page work pays off.

Step 4: Interpret the baseline and prioritize

Read results along three axes: visibility, accuracy, and sentiment. Strong map-pack reach with low AI mention signals a GEO gap. High mention with low accuracy is a trust and data problem. Negative framing despite correct core data points to review and reputation work. Prioritize fixes by impact: eligibility and data consistency first, then citable local content formats, then distribution through trusted third-party sources.

Repeat the same prompt set quarterly under identical location and session conditions. Only then does a snapshot become a trend. Also document which measures were implemented between runs, such as NAP cleanup, new location pages, or stronger presence on review platforms. The local GEO baseline audit does not replace classic local SEO; it adds the question of whether generative systems recommend the business, represent it correctly, or favor competitors — before more budget goes into content and citations.