Skip to content
GeoLumen

How to show up in Claude: a small-business guide

When Claude searches the web, it relies on a search partner's index and Anthropic's own crawlers. Be readable by both.

Updated 21 Oct 2026 · 3 official sources

How does Claude find businesses?

Claude answers from its training data, and, when web search is on, from live search results with citations. Anthropic lists Brave Search among its subprocessors, and independent tests found Claude's search citations matching Brave's results. Anthropic's help center also refers to "our partners" indexing content for Claude's web search.

Anthropic runs three crawlers with separate jobs, and you can control each in robots.txt. Brave's crawler doesn't announce its own user agent; Brave says that if a page isn't crawlable by Googlebot, Brave won't crawl it either.

Which crawlers matter for Claude?

Claude crawlers and what to do with them
User agentDo
Claude-SearchBot
Search

Analyses content to improve Claude's search results. Anthropic says blocking it may reduce your visibility in Claude's answers.

Allow
Claude-User
User-triggered

Fetches pages when a Claude user asks a question that needs them.

Allow
ClaudeBot
Training

Collects public content that may be used to train Claude. Blocking it signals your future content should be excluded from training.

Your choice

robots.txt you can copy

User-agent: Claude-SearchBot
Allow: /

User-agent: Claude-User
Allow: /

# Optional: opt out of model training only
# User-agent: ClaudeBot
# Disallow: /

Not sure what your file says now? Check it free: it reads your robots.txt rule by rule for Claude-SearchBot, Claude-User, ClaudeBot and 18 other AI bots.

What should you fix, step by step?

  1. 01Allow Claude-SearchBot and Claude-User

    Anthropic says its bots honour robots.txt (including Crawl-delay) and don't bypass CAPTCHAs, so a robots.txt block or a challenge page will keep them out. Check your file with the AI Crawler Checker, and make sure your CDN's bot protection isn't challenging them.

    Rules apply per host: if your site lives on www. and a booking subdomain, each needs its own robots.txt.

  2. 02Make sure Brave Search has your pages

    Search for your business name and your main services at search.brave.com. Brave has no webmaster console or sitemap submission, but you can ask it to re-fetch a page with its "Submit URL" form. Since Brave follows Googlebot's permissions, a page Googlebot can't crawl won't be in Brave either.

    Pages that other sites link to get discovered faster, which is another reason to be listed in directories and local press.

  3. 03Put your key facts in plain HTML text

    Claude-SearchBot reads the HTML your server sends. Facts that only exist inside images, PDFs, sliders or text loaded later by JavaScript are easy to miss. Your homepage and service pages should say, in ordinary text: what you do, for whom, where (city and areas served), how to book, hours, and typical prices if you can share them.

    One page per main service, with the questions customers actually ask answered on it, gives an assistant something specific to quote. Run the SEO Health Snapshot for the basics, and see AI crawlers and JavaScript.

  4. 04Add LocalBusiness schema that matches your listings

    Structured data (JSON-LD) states your facts in a form machines don't have to interpret: business type, address, phone, opening hours, area served, price range and your profile links in sameAs. It doesn't guarantee a mention, but it removes ambiguity, and it must say exactly what your listings say.

    Use the most specific type that fits (Dentist, Plumber, Attorney…). Our free Schema / JSON-LD Generator writes it for you.

  5. 05Make your name, address and phone identical everywhere

    AI answers are stitched together from many sources at once: your site, your listings, review sites and directories. When they disagree (an old phone number, a previous address, "St." on one and "Street" on another plus a different suite), the assistant has to guess, and it often picks the version that appears most.

    Write down one canonical version of your business name, address, phone, hours and website, then make every profile match it: Google Business Profile, Bing Places, Apple Business, Yelp, Facebook, and the directories for your trade.

  6. 06Earn real reviews where this engine looks

    When someone asks for "the best" or "a trusted" business, assistants lean on review platforms and third-party lists rather than your own claims. Claude's citations for local and "best" questions come from the same open web: review platforms, trade directories and articles. Check which ones Claude cites for your category.

    Ask happy customers for reviews as part of your normal process, reply to reviews (good and bad) and fix the issues they raise. Never buy or script reviews: platforms remove them, and it can get a listing suspended.

  7. 07Check what Claude says about you, then re-check monthly

    Ask Claude the questions your customers ask ("best emergency plumber in Leeds", "dentist open Saturdays near Austin"), not just your business name. Note who gets named, what's said about you, and which sources are cited. Make sure web search is enabled in Claude's settings, or you'll only see what the model remembers from training.

    Our free AI Visibility Quick Scan does a first pass, and this guide shows how to test by hand.

  8. 08Optional: add an llms.txt

    llms.txt is a proposed plain-text summary of your site for AI tools. None of the major assistants has said it uses llms.txt to choose what to cite, so treat it as a cheap extra, not a fix. It takes ten minutes with our llms.txt Generator. More in Does llms.txt actually help?

Sources

We update this guide when Claude changes how it works. Something out of date? Tell us.

FAQ

Claude: common questions

Rather not do this yourself? We'll do the whole thing and send you the receipts.