For twenty years the job of getting found was simple to describe: rank on Google. That world is ending. People now open an assistant, describe what they need in a full sentence, and accept the single answer it hands back. The business that is that answer wins the customer before a list ever loads.
We have spent the last year watching which businesses get named when someone asks ChatGPT, Gemini, or Perplexity for a recommendation — and which get skipped. The pattern is consistent enough to write down. Here is the five-step playbook, in the order we run it.
In an AI answer there is no page two. There is one recommendation, and maybe a runner-up.
Key idea
Claim your machine-readable storefront
Your homepage was built for people who scroll. Assistants do not scroll — they retrieve. The first move is publishing a second, structured version of your business: hours, services, pricing bands, service area, and proof, laid out so a model can lift a clean answer in one pass instead of guessing from marketing copy.
Structure facts the way models retrieve them
Bury a fact inside a paragraph and it competes with everything around it. Surface it as a labelled pair — "Emergency visits: same-day, 7 days a week" — and it becomes quotable. Every claim you want repeated should be short, attributable, and sitting next to the question it answers.
Answer the questions people actually ask
Nobody types "premier dental solutions." They ask "is there a dentist near me open on Sunday that takes my insurance?" Map the real, messy questions your customers ask an assistant, then write the answer plainly. The closer your wording is to the question, the more often you are the response.
Earn citations, not just clicks
In an AI answer there is no page two — there is one recommendation and maybe a runner-up. Being cited means being consistent everywhere a model looks: your storefront, your listings, your reviews all telling the same story. Contradictions are the fastest way to get left out.
Monitor, measure, and feed the loop
You cannot improve what you cannot see. Track which assistants mention you, for which prompts, and how accurately. Every gap is a content task; every win is a pattern to repeat. Discovery is not a launch — it is a loop that compounds the longer you run it.
Start before your competitors notice
The shift to AI discovery is early enough that the businesses moving now are setting the defaults. Claim your structured storefront, write plainly, stay consistent, and watch the loop. Six months from today, "we get recommended by AI" will sound ordinary. Right now it is still an edge.
