An adult daughter looking for memory care for her father no longer opens ten browser tabs. She opens ChatGPT and types “best memory care near Austin for a parent with early dementia.” She gets one clean answer that names three communities and explains why.
Yours isn’t one of them. That conversation ended before anyone on your team ever knew it happened — and AI is increasingly the front door: families ask before they click. So the question every operator is now asking is the right one: how do you get recommended by ChatGPT? Not with a hack, and not by buying a rank that doesn’t exist. You raise a visibility rate with a repeatable loop.
Start with your baseline. Run the free AI-visibility check → and see, per engine, how often ChatGPT, Google AI Overviews, and Perplexity actually name your community today.
First, the honest framing: you’re raising a rate, not buying a rank
Before any tactics, internalize the one thing most vendors won’t tell you: AI answers are non-deterministic. Ask ChatGPT the same family question twice and you can get two different lists of communities. There is no “#1 in ChatGPT” to purchase, and anyone selling a guaranteed “AI rank” is selling a number that doesn’t exist.
Why it matters: the only thing you can honestly move — and measure — is your visibility rate: how often you’re named across many runs of the questions families actually ask. The right way to report it isn’t a rank; it’s a rate with a confidence interval, like “named in 6 of 10 ChatGPT answers (60%, 95% CI 31–83%).”
And you must measure it per engine. “AI” is not one thing. ChatGPT, Google AI Overviews, Gemini, and Perplexity each assemble answers from different sources, so strong visibility in one tells you nothing about the others. The loop below is built on that reality.
Step 1 — Measure your baseline (per engine)
You cannot improve what you cannot see, and a single screenshot proves nothing because the next run may differ. A real baseline means asking the questions families ask — “best assisted living in your city,” “memory care near your city for a parent with dementia” — repeatedly, across every engine, and recording how often you appear.
What to do: establish a per-engine visibility rate before you change anything. That’s your before-picture, and without it you’ll never prove the work paid off. This is exactly what the free checker does — it repeats the queries across engines and reports a rate with a confidence interval, plus which competitors get named instead and which sources the engines cite.
Why it matters: the baseline tells you which engine is leaking families. If Perplexity names you 70% of the time but ChatGPT never does, your problem isn’t “AI” — it’s the specific sources ChatGPT reads. That diagnosis is what makes Step 2 efficient instead of guesswork.
Step 2 — Fix the sources each engine actually reads
Once you know where you’re invisible, fix the inputs that engine pulls from. AI engines don’t invent facts about your community; they repeat the most authoritative source they can find. Give each engine something clear and consistent to repeat.
| Engine | Pulls heavily from | What to fix first |
|---|---|---|
| ChatGPT / Copilot | Bing’s index, Bing Places, Yelp, BBB | Claim and fully complete Bing Places (not just Google); fix Yelp and BBB listings |
| Google AI Overviews | Google Business Profile, Google reviews, Reddit | Complete GBP, build a steady, answered-review motion, maintain an authentic Reddit presence |
| Perplexity | Cited, crawlable web content with clear sourcing | Publish factual, well-sourced pages that are easy to quote and attribute |
| Gemini | Google’s ecosystem + the broader web | Overlaps with Google AIO; keep facts consistent across the web |
The most common mistake: an operator completed their Google profile years ago and never touched Bing — but ChatGPT and Copilot lean on Bing, so they’re invisible on the engine the family happened to use. Claim Bing Places, make name/address/phone/care-types/hours identical across every listing, and put your facts in plain text, not in a brochure PDF or an amenities infographic — AI can’t read an image. Then earn and respond to reviews: volume and recency are heavy citation signals, especially for Google AI Overviews.
Watch the aggregators here. When an engine can’t confidently describe you, it reaches for a source it trusts — often A Place for Mom, Caring.com, or Seniorly — and recommends “contact A Place for Mom” instead of naming you. This isn’t hypothetical: Senior Housing News reported in late 2025 that A Place for Mom is reorienting its whole strategy around AI. Every move-in they intermediate costs you a referral fee of roughly 50–120% of the first month’s rent — commonly $3,500–$12,000 per move-in. Being the source the engine repeats is how you keep that fee.
Re-check after each fix. Run the free AI-visibility check → again once your listings and reviews are squared away — the per-engine rate is the only proof a change actually landed.
Step 3 — Become a citable source in your own right (this is GEO)
Fixing listings and reviews gets you described. To get recommended — consistently, across engines — you have to become something the AI wants to quote directly. That discipline has a name: Generative Engine Optimization (GEO).
What to do: publish clear, factual, well-structured pages about your community and your market — care types, pricing approach, neighborhood, “who we’re the right fit for,” current availability — written as concise, quotable text with comparison tables, direct-answer FAQ blocks, and schema markup. The goal is for an engine answering “best assisted living in your city” to be able to cite you rather than an aggregator.
Why it matters: listings make you accurate; citable content makes you the answer. This is the highest-leverage, most durable move, and it’s the heart of the pillar guide — see GEO for senior living for the full per-engine playbook.
Step 4 — Re-measure to prove the rate moved
A fix you can’t measure is a story, not a result. After the work in Steps 2 and 3, run the same queries across the same engines and compare the new visibility rate to your baseline.
What to do: treat this as a standing loop — measure → improve → re-measure — on a regular cadence, not a one-time project. AI sources shift, competitors invest, and your rate drifts; the operators who win are the ones who keep score.
| Loop stage | The question it answers | The honest metric |
|---|---|---|
| Measure | ”Where do we stand, per engine?” | Baseline visibility rate + confidence interval |
| Improve | ”What sources is each engine missing?” | Listings, reviews, citable plain-text pages |
| Re-measure | ”Did the work move the rate?” | Before/after rate change, per engine |
Why it matters: because answers are non-deterministic, the before/after comparison — not a vendor’s promise — is your only honest proof. If the rate moved up on the engine you targeted, the loop worked. If it didn’t, you’ve learned exactly where to push next, instead of guessing.
The part that actually matters: does it drive move-ins?
Getting named by ChatGPT is worth nothing if you can’t tell whether it produced a paying resident. Senior living is uniquely hard to attribute: a 107-to-400-day decision cycle, an adult child doing the research while the parent moves in, and a “how did you hear about us?” form answered months after the AI conversation that started everything — by which point the family barely remembers, which is why that question is unreliable.
So raising your AI visibility only pays off when it’s joined to closed-loop attribution — and the honest version is directional, with a match confidence, never a deterministic click trail. You can’t trace one ChatGPT conversation deterministically to one move-in, but you can connect AI-sourced inquiries to tours to move-ins with a stated confidence, turning a vanity metric into proven occupancy. (See the complete attribution guide and why “how did you hear about us?” can’t be trusted.)