Make expertise explicit.
Services, identity and contact information should be consistent. Each piece answers a defined question with specific examples, limitations and sources where appropriate. A helpful short answer can be followed by a detailed explanation without artificially padding the content.
Make content accessible.
Readable HTML, useful internal links and correct indexability are foundations. Structured data describes what is actually visible. An llms.txt file can provide an editorial guide, but it is neither a universal requirement nor a guarantee a model will use it.
Observe carefully.
Generated answers vary with the system, prompt and time. We document a question set and track observed mentions without presenting those observations as a stable ranking. The goal remains reliable content that also helps human readers.