In short
- The website must describe services, audiences and regions unambiguously.
- Sources and contact paths should be publicly verifiable.
- Agent functions need clear names, rules and internal recipients.
1. Content clarity
First check whether a person understands in less than a minute what your business offers, who it's for and which region it serves.
If this information sits only in images, sliders or generic claims, it's often too weak for AI systems.
2. Sources and evidence
For each important service there should be a stable public source. That can be a service page, documentation, contact page, calculator page or API description.
Sources help AI systems distinguish verified facts from marketing claims.
3. Contact and functions
A profile becomes AI-agent-ready only when an agent can not just read but also recognize an allowed action.
Typical functions are request a rate, request a quote, request a meeting, request a callback or ask a question. Each function should be routed internally to the right person.
4. Crawler access: let OAI-SearchBot and Claude-SearchBot read your site
agentready.ch profiles are open to every crawler through robots.txt, so ChatGPT, Claude and other AI systems can read them; for your own website, check the same user agents in robots.txt and in your firewall or CDN.
OpenAI uses three agents. OAI-SearchBot builds the index for ChatGPT search, and sites that block it are not shown in ChatGPT search answers. GPTBot collects content that may be used for training. ChatGPT-User fetches a page when a person asks ChatGPT to, and because a user starts that visit, robots.txt rules may not apply. Details: OpenAI crawler documentation and Searching the web with ChatGPT.
Anthropic also uses three. Claude-SearchBot indexes pages for search results, Claude-User fetches pages when a Claude user asks a question, and ClaudeBot collects content that may be used for training. All three follow robots.txt. Details: Anthropic on its crawlers.
Perplexity shows and links websites in its search results with PerplexityBot; Perplexity says it does not use it to crawl content for AI foundation models. Perplexity-User fetches a page when someone asks Perplexity a question and generally ignores robots.txt. Details: Perplexity on its crawlers.
Google and Apple read with their search crawlers, Googlebot and Applebot. For training, both have a separate token without a crawler of its own. Google-Extended manages whether Google may use content for training Gemini and for answers in the Gemini apps; it does not affect Google Search. Disallowing Applebot-Extended keeps your content out of training Apple's foundation models; the pages can still appear in search results. Details: Google on its crawlers and Apple on Applebot.
Search and training are separate settings. You can let the search agents read your site and still keep your content out of model training, as in the example below. Disallowing Google-Extended as well also keeps your content out of answers in the Gemini apps. agentready.ch itself uses User-agent: * with Allow: / and blocks none of these crawlers.
A firewall or bot protection can block these agents even when robots.txt allows them. OpenAI asks hosts and CDNs to allow its published crawler IP ranges, so check your security settings as well.
robots.txt: allow AI search, opt out of training (optional)
# Let AI search and user-triggered visits read the siteUser-agent: OAI-SearchBotAllow: / User-agent: ChatGPT-UserAllow: / User-agent: Claude-SearchBotAllow: / User-agent: Claude-UserAllow: / User-agent: PerplexityBotAllow: / User-agent: Perplexity-UserAllow: / # Optional: keep content out of model trainingUser-agent: GPTBotDisallow: / User-agent: ClaudeBotDisallow: / User-agent: Applebot-ExtendedDisallow: /Checklist
- Describe the main service in one clear sentence
- Allow OAI-SearchBot, Claude-SearchBot and PerplexityBot in robots.txt and in your firewall or CDN
- Decide on training (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended) deliberately and separately from search
- Make region, language and audience visible
- Link service and contact pages publicly
- Provide agent.md and agent.json
- Define at least one allowed agent function
- Set internal recipients per function
- Run a dry-run test via Agent API or MCP
Profiles in the directory
These companies have an agent-ready profile on agentready.ch. Each profile shows services, region, sources and the contact routes AI agents may use.
ThatDay
Insurance and advisory · Switzerland
PrimAI
Swiss Health Insurance Intelligence · Switzerland
Expat Savvy
Insurance and pension advisory · Zurich
Welcome Service
Relocation & welcome services · Switzerland (Suisse Romande + CH)
Unterwaldner
Insurance & pension advisory · Central Switzerland (OW/NW/LU/UR)
3 Grad Physiotherapie GmbH
Physiotherapy · Winterthur
Frequently asked
Does a business need its own API?
No. To start, email-based functions through agentready.ch are often enough. Your own API matters only when calculators, bookings or process data should be connected directly.
How fast can a base profile be created?
A profile can be prepared from public sources. A verified profile additionally needs a personal review and clear contact rules.