To get your brand mentioned in Claude, make your site readable by Anthropic's search and user-fetch crawlers, rank well in the search results Claude pulls from when web search is on, and build a consistent, well-documented footprint across the web so the model already knows you when it answers from training alone. Claude differs from consumer-first assistants because so much of its use happens at work: it is heavily used by businesses, runs inside tools like Microsoft 365 Copilot, Excel and Slack, and when it searches it shows its sources inline.

That makes Claude a B2B channel first. If you sell software, services or anything researched by professionals, this guide covers how Claude chooses brands and what to do about it.

Quick takeaways

  • Anthropic runs three separate crawlers: ClaudeBot (training), Claude-User (user-requested fetches) and Claude-SearchBot (search quality and indexing).

  • Anthropic says all three honor robots.txt, and blocking ClaudeBot does not stop the search or user bots.

  • Reporting and Anthropic's subprocessor list point to Brave Search as the main web search backend; Anthropic's help docs do not name a provider.

  • Claude's help center says every web search response includes citations, so being the cited page matters as much as being mentioned.

  • Anthropic has pledged to keep Claude ad-free, so there is no paid route into its answers.

  • In Team and Enterprise workspaces, an admin must switch web search on, so many work answers still come from training data.

  • Each Claude model has a fixed knowledge cutoff; anything newer only reaches Claude through search or user-supplied content.

How Claude decides which brands to mention

Training data: what Claude already knows

Without search, Claude answers from what it learned during training. Anthropic's help center explains that each model has a fixed knowledge cutoff and may not know about events after it; the current 5.5 generation has a reliable knowledge cutoff of June 2026. Training data includes public web content collected by ClaudeBot. If your brand was clearly and repeatedly described in reviews, comparisons, documentation and news before that cutoff, Claude can recommend you from memory. If you launched or rebranded afterwards, Claude only knows you through search.

Web search: what Claude retrieves live

Anthropic introduced web search in March 2025 and extended it to every plan worldwide in May 2025. Claude decides when to search, may run several queries for one question, and links its sources directly in the answer. Anthropic's web search help page says every response includes citations, with direct links and relevant quotes where appropriate.

Anthropic does not name its web search provider in its help docs. However, TechCrunch reported in March 2025 that Brave Search was added to Anthropic's subprocessor list the week search launched, and Simon Willison found a BraveSearchParams parameter in Claude's search function plus identical citations between Claude and Brave for a test query. The same help page says image results come from Bing. For brand visibility, the practical reading is: your rankings in Brave Search matter for Claude in a way they do not for most other assistants.

Where search is switched off

This is the detail many teams miss. In Claude Team and Enterprise plans, an owner must enable web search for the whole organization before members can use it, according to the help center. Developers using the Claude API web search tool choose whether to enable search at all, can restrict it with allowed or blocked domain lists, and can localize results by country or city. So the same question can be answered from training data in one company, from live search in another, and from a filtered source list inside a third-party app.

In short: Claude mentions brands from its training data by default and from Brave-backed live search when search is available, and in business settings search is often off, so long-term reputation counts more here than on most engines.

Why Claude matters for B2B brands

Claude has a strong enterprise footprint. Menlo Ventures data reported by TechCrunch in July 2025 put Anthropic at 32% of enterprise LLM usage versus 25% for OpenAI, and in April 2026 PYMNTS reported more than 1,000 business customers paying over $1 million a year. Claude is also embedded in workflows: Anthropic lists Claude in Chrome, Excel, PowerPoint and Slack plus Amazon Bedrock and Google Cloud Vertex AI, and Microsoft added Claude to the Researcher agent and Copilot Studio in Microsoft 365 Copilot in September 2025. Buyers researching vendors inside these tools may never open a search engine.

What Claude tends to cite

Claude favors pages that answer the exact question in clean, quotable passages: documentation, comparison articles, analyst-style reviews and community discussions. KIME research across ChatGPT, Perplexity, Gemini and Claude found Reddit in roughly 1 in 4 AI answers, YouTube just under, and a top 5 of Reddit, YouTube, Wikipedia, LinkedIn and Trustpilot, with no brand-owned sites in that top 5. Use KIME citation data to see which domains Claude specifically cites for your category, because the mix differs by engine.

Anthropic crawlers and what they control

Anthropic documents three user agents in its crawler help article (updated April 2026). Search Engine Journal notes that, unlike OpenAI's user-triggered fetcher, Anthropic says its user bot honors robots.txt too.

User agent

What it does

What blocking it means

ClaudeBot

Collects public web content that may be used for model training.

Future content is excluded from training. Search and user fetches are unaffected.

Claude-SearchBot

Crawls to improve search result quality and relevance.

Your content is not indexed for search, which may reduce visibility in Claude's answers.

Claude-User

Fetches pages when a user's question requires it.

Claude cannot retrieve your page for users, reducing visibility in user-directed search.

Anthropic also supports the Crawl-delay directive and advises against IP blocking, which may not reliably opt you out and can stop its bots from reading robots.txt.

How to get your brand mentioned in Claude: 8 steps

1. Allow Claude-SearchBot and Claude-User

Check robots.txt on every subdomain, plus CDN and bot-management rules, for Claude-SearchBot and Claude-User. Blanket AI-bot rules often catch these by accident. Decide separately whether to allow ClaudeBot: allowing it helps future models know your brand; blocking it does not hurt search. One KIME customer found an AI crawler blocked from their site entirely, and unblocking it made an immediate measurable difference. KIME AI Site Health flags these issues automatically.

2. Check how you rank in Brave Search

Run your key category and comparison queries in Brave Search and see where you and your competitors land. Brave runs its own independent index, so pages that rank in Google are not guaranteed to appear. Make sure your important pages are crawlable, internally linked and fast, and that your sitemap is clean.

3. Write for citation, not just for clicks

Claude quotes short passages, so give it quotable ones. Lead each section with a direct answer, use question-style headings, state facts with numbers and dates, and put comparison data in HTML tables. KIME's guide to structuring content for LLM extraction explains the formats that survive retrieval, and FAQ schema helps map common questions to your answers.

4. Publish documentation-grade content

Claude is used heavily for technical and professional work, and it leans on clear reference material. Public docs, pricing pages, security and compliance pages, integration guides and changelogs give Claude precise facts to cite when a buyer asks whether your product supports SSO, which regions you host in or how you compare on price. Keep them public, indexable and current.

5. Build third-party coverage where professionals look

Because many enterprise answers come from training data, your reputation across the web over time decides whether Claude knows you. Earn mentions in industry publications, analyst-style comparisons, review platforms such as G2, Capterra or Trustpilot, LinkedIn thought leadership, YouTube walkthroughs and genuine Reddit or community discussions. See how to get cited by AI search for tactics.

6. Make your entity consistent

Use one brand name, one short description and the same category language on your site, LinkedIn, Crunchbase, Wikipedia or Wikidata where eligible, and review profiles. Add Organization schema with sameAs links. Consistency helps Claude connect scattered mentions to a single entity, which matters most when it answers from memory.

7. Prepare for agents, not just chat

Claude increasingly acts on the web through Claude in Chrome and connected tools. Pages with clear navigation, accessible forms and server-rendered content are easier for agents to read and use. KIME's checklist on whether your site is agent-ready covers the basics.

8. Measure, fix and repeat

Track a fixed set of prompts over time so you can tell whether changes worked. Look at which URLs Claude cites for competitors and target those pages first. KIME Action Center turns visibility gaps into a prioritized task list.

In short: let Anthropic's search bots in, rank in Brave, and build the kind of documented, widely cited reputation that Claude can recall even when search is off.

How to check if Claude mentions your brand

Start with a structured manual test:

  • Write 10 to 20 prompts your buyers would ask, including category questions, comparisons with named competitors and problem-led questions with no brand in them.

  • Run each prompt with web search on and again with it off, to separate training-data visibility from search visibility.

  • Repeat each prompt three to five times, since answers vary between runs.

  • Note the model you tested, because different Claude models have different knowledge cutoffs.

  • Record mentions, position, sentiment and the exact URLs cited in each answer.

Doing this weekly across markets is slow. KIME tracks Claude alongside nine other engines, running your prompts by category and market and reporting visibility, placement, share of voice, sentiment and the domains and URLs Claude cites. You can track your Claude visibility with KIME, or compare options in KIME's overview of AI visibility tools.

Common mistakes

  • Blocking all Anthropic bots to stop training, which also removes you from Claude's search and user fetches.

  • Checking only Google rankings and never looking at Brave Search.

  • Testing with web search on only, and missing that many workplace answers come from training data.

  • Hiding pricing, documentation or integration details behind logins or forms that Claude cannot read.

  • Assuming Claude sees your latest launch without fresh, crawlable coverage of it.

Frequently asked questions

Does Claude use Google or Bing for web search?

Claude does not appear to rely on Google or Bing for its main web results. Anthropic's help docs do not name a provider, but reporting in 2025 found Brave Search on Anthropic's subprocessor list and Brave-specific parameters in Claude's search tool. Anthropic's help center does say image results come from Bing. Checking your Brave Search rankings is a sensible step.

Can you pay to be mentioned in Claude?

You cannot pay to be mentioned in Claude's answers. Anthropic pledged in February 2026 that Claude will stay ad-free and that advertisers will not influence its responses, so mentions come from training data and live search results. The way in is earned visibility: crawlable pages, strong rankings in the search results Claude retrieves, and credible third-party coverage that Claude can recall or cite.

Why doesn't Claude mention my brand?

Claude usually leaves a brand out because it cannot crawl the brand's pages, the brand does not rank in the search results Claude uses, or the brand had little coverage before the model's knowledge cutoff. Check robots.txt for Claude-SearchBot and Claude-User, test Brave Search rankings, and build consistent third-party mentions.

Should I block ClaudeBot?

Blocking ClaudeBot is a choice about training data, not search visibility. Anthropic says blocking it excludes your future content from training, while Claude-SearchBot and Claude-User work independently. The trade-off is that future Claude models may know less about your brand when answering without web search, which is common in business workspaces.

How long does it take to get mentioned in Claude?

Getting mentioned in Claude through web search can happen within days or weeks once your pages are crawlable and rank in Brave Search. Getting into Claude's built-in knowledge takes longer, because it depends on new model releases trained on data collected after your coverage appeared. Plan for both timelines.

Does Claude cite its sources?

Claude cites its sources whenever it uses web search. Anthropic's help center says every web search response includes citations, with links and relevant quotes, and the API returns cited text with each source URL. Answers given from training data alone do not include citations, so test both modes when checking visibility.

Is Claude worth optimizing for if most people use ChatGPT?

Claude is worth optimizing for if you sell to businesses or professionals. Menlo Ventures data reported in 2025 showed Anthropic leading enterprise LLM usage, and Claude runs inside Microsoft 365 Copilot, Excel, Slack and cloud platforms. Many of the steps also improve visibility in other AI engines, so the effort is rarely wasted.

Does Claude know about recent product launches?

Claude knows about recent product launches only if web search is on and can find coverage of them. Each Claude model has a fixed knowledge cutoff, and anything newer reaches Claude through search or content a user shares. Publish launch details on crawlable pages and secure third-party coverage quickly.

Related guides

Sources: Does Anthropic crawl data from the web, and how can site owners block the crawler? · Anthropic's Claude Bots Make Robots.txt Decisions More Granular · Claude can now search the web · Enabling and using web search · Web search tool · Anthropic appears to be using Brave to power web search for its Claude chatbot · Anthropic use Brave · How up-to-date is Claude's training data? · Enterprises prefer Anthropic's AI models over anyone else's, including OpenAI's · Anthropic Hits $30 Billion Run Rate as Enterprise Demand Accelerates · Claude is now available in Microsoft 365 Copilot · Anthropic overview · Anthropic cements its position as the not-OpenAI with no-ads pledge

Picture of Nikolaj Reissmann, Co-founder and COO at KIME

Nikolaj Reissmann

Co-founder & COO

Share