STUDIO NOTES / SEARCH & DIGITAL PRESENCE
AI Search Readiness: Access, Identity and Evidence
AI search readiness is a review of the information available about your business and the systems that can discover it. It is not a promise that a particular assistant will recommend you.
Explain the business before chasing a prompt
A useful business description says who you are, what you offer and where you work. For Befound Studio, that means a Bloemfontein base, a South African service area and clear descriptions of brand management, web design, Google Ads and the studio’s creative services.
Those facts should agree across the website, contact details and verified public profiles. Structured data can connect the pages to the same organization, but it should not introduce claims that visitors cannot verify on the site.
Check access and indexing separately
Review public HTML, meaningful links, robots instructions, sitemap entries and canonical URLs. Then use available search-owner tools to investigate indexing and visibility. A page returning HTTP 200 is accessible; it is not evidence of indexing or an AI citation.
Google also provides a Search generative AI control in Search Console. Its current default is inclusion, but an owner should check the setting for the actual property and any inherited configuration.
Search crawling is not the same as training
Providers have different controls. OpenAI distinguishes OAI-SearchBot, used for search discovery, from GPTBot, used for model training. A business can make a separate training choice without assuming that training access is required for search visibility.
Provider documentation and the business’s preferences should guide these decisions. Do not copy a universal “allow every AI bot” or “block every AI bot” rule without understanding its purpose.
Give people useful evidence
A case study is more useful when it identifies the actual contribution, the problem addressed and what can be verified. Separate delivered work from proposed mockups. Only publish conversion, revenue or other performance improvements when the underlying evidence supports them.
The same principle applies to service pages. Explain deliverables, scope boundaries, required access and reporting. Practical answers based on real work are more valuable than repeating the same broad claims across many pages.
Measure without promising placement
Establish a baseline using the tools available, such as Search Console’s search and generative AI reports or Bing Webmaster Tools’ AI Performance. A manual prompt test is a snapshot: answers, sources and interfaces can change. Record what was tested instead of presenting a single answer as a permanent ranking.
Google says llms.txt is not used for its Search visibility. Useful content, accurate identity and sound technical foundations should come before supposed AI shortcuts.