Local Business Visibility in Claude

Local businesses appear in Claude's recommendations through the open web, not through a maps product. Claude has no native business-listing system, so when a user asks for "the best coffee roaster in Portland," it retrieves web pages — review sites, local roundups, directories, and your own site — through its Brave Search-backed index and synthesizes an answer from what those sources agree on. Winning means dominating the written record about your business.

How does Claude answer local recommendation prompts?

When a local-intent prompt arrives, Claude's web search tool (rolled out in March 2025) issues queries against its index and fetches candidate pages live with Claude-User. It then favors businesses that appear across multiple corroborating sources. A restaurant mentioned on its own site, a Yelp profile, an Eater roundup, and a local subreddit gets recommended; one that exists only on its own site usually does not. Consensus beats assertion — this is the single most important mechanic to internalize.

Claude also cannot see anything behind JavaScript. Its fetchers take raw HTML without rendering, so a site built as a client-side app shell is invisible at fetch time. Server-render your homepage, location pages, and menu or service pages.

The local visibility stack for Claude

Why NAP consistency still matters in an AI answer engine

Claude resolves your business as an entity by matching facts across sources. If your address reads "123 Main St, Suite 4" on your site, "123 Main Street" on Yelp, and an old address on a stale directory, the model gets conflicting evidence and its confidence in recommending you drops. Audit every listing once a quarter; fix the ones that disagree. The schema.org LocalBusiness type — with address, geo, telephone, and openingHoursSpecification properties — gives Claude a machine-readable version of the same facts on your own domain.

What content should a local business publish?

Location-specific answer pages outperform generic brochureware. A page titled "Wedding photography pricing in Denver (2026)" with real numbers is quotable; a homepage that says "capturing your special moments" is not. Publish the pages that answer what locals actually ask: pricing, hours, comparisons with nearby alternatives, seasonal availability, parking. Each answer should be a self-contained 40-80 word passage, because that is the unit Claude extracts. The same structure that wins here also feeds AEO across other engines.

How do you track local visibility in Claude?

Build a fixed prompt set of 15-25 geo-specific prompts — every service you offer crossed with every neighborhood you serve — and run it on a weekly cadence, logging mentions, cited URLs, and which competitors appear. Manual checking works at small scale; AI visibility tracking automates the sampling and trends your share of local recommendations over time, including which review platforms Claude is actually citing in your market. That last data point tells you exactly where to invest your next ten hours of profile work.