How an AI answer about senior living is built
When a family asks an assistant something like "what are good memory care options near San Mateo," assistants that browse the web run searches, read a handful of pages, and write a summary. Some also draw on what the model learned in training. Google AI Overviews combine Google's index with a generated summary.
Across these systems, the same pattern holds: the assistant can only recommend what it can find, read, and trust. That gives operators three levers: be findable, be readable, and be corroborated.
Be findable
- Allow search and AI crawlers in robots.txt, including GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, and Google-Extended, unless you have a reason not to.
- Rank in regular search for your care types and city. Browsing assistants lean on search results.
- Keep Google Business Profile, Apple, and Bing listings complete and accurate.
- Publish an llms.txt file that summarizes your community and links to your key pages.
Be readable
Assistants quote passages that answer a question completely in a few sentences. Pages that bury the answer under slogans or load their text with scripts are hard to use.
- Open each page with a two to four sentence answer.
- Use question-style headings that match what families ask.
- Serve text in the HTML itself so crawlers do not need to run scripts.
- Add JSON-LD structured data for your organization, each location, services, and FAQs.
- State specifics: care types, licensed capacity, neighborhoods served, how to tour, and a pricing range where you can.
Be corroborated
Models are more confident about facts that appear consistently across independent sources. A community listed under three slightly different names with two phone numbers is less likely to be named.
- Use one exact name, address, and phone number everywhere.
- Keep directory and referral site profiles current.
- Grow authentic reviews and reply to them.
- Earn local mentions from news, partners, and associations.
Measure it
Build a fixed set of realistic family questions for your markets and care types, run it across assistants monthly, and record whether you are mentioned, whether you are cited, and whether the details are right. Track the same set over time so changes are meaningful.