A family three states away opens ChatGPT and types “best assisted living near
Now run the same scenario through the channel you’re probably paying for. That same family lands on an aggregator, fills out a form, and gets routed to whoever’s paying for the lead. When they move in, you write a check for somewhere between half and all of their first month’s rent — commonly $3,500 to $12,000. The work of earning the family was identical. The difference is who got named first, and whether it cost you anything.
Don’t guess — check. Run the free AI-visibility check → and see, per engine, how often AI actually names your community when families ask.
This is a playbook for shifting your inquiry mix toward direct. Not eliminating aggregators overnight — they still drive real volume — but cutting your dependence on them by earning the inquiries you’d otherwise pay for, and proving the shift with attribution you can defend in a budget review.
Lever 1 — AI visibility is the new direct channel
What it is: Families increasingly ask ChatGPT, Perplexity, Gemini, and Google’s AI Overviews for shortlists before they ever touch an aggregator or a search results page. The community the AI names gets the inquiry direct — at no referral cost. This is the same disintermediation that made aggregators powerful in the first place, except this time the front door can be yours. It’s not hypothetical for the category: Senior Housing News reported in late 2025 that A Place for Mom is reorienting its business around AI. When the largest aggregator is repositioning around the AI front door, that’s a signal about where families are starting.
Why it’s a direct channel: When an aggregator names you, the family enters their funnel and you pay on move-in. When ChatGPT names you, the family enters your funnel. Same demand, different owner of the relationship — and a very different cost structure. For the full economics, see what aggregators really cost.
How to do it — honestly: You cannot pay to “rank #1” in AI answers. The answers are non-deterministic; ask the same question twice and you may get different communities. So the goal isn’t a rank — it’s a visibility rate reported with a confidence interval, measured per engine: out of N realistic family questions, how often does each engine name you, and how confident are we in that number. ChatGPT might name you 40% of the time while Google AI Overviews names you 5%. Those are different problems with different fixes, which is why you measure each engine separately rather than chasing one blended “score.”
The levers below are how you raise that rate. They’re also just good fundamentals — they help families and human searchers too — which is what makes them durable rather than a trick that stops working next model update.
Lever 2 — Complete, consistent listings on the sources each engine reads
What it is: AI engines don’t invent facts about your community; they assemble them from sources they trust. If those sources are thin, stale, or contradict each other, you’re hard to name confidently — and a hesitant engine names a competitor instead.
The catch most operators miss: Different engines read different sources, so a single “set up our Google profile” project leaves gaps. Map your work to where each engine actually looks:
| Engine | Reads from (primary sources) | What to fix first |
|---|---|---|
| ChatGPT / Copilot | Bing, Bing Places, Yelp, BBB | Claim and complete Bing Places; fix Yelp and BBB listings |
| Google AI Overviews | Google Business Profile, Google reviews, Reddit | Complete every GBP field; keep reviews fresh; watch Reddit threads |
| Perplexity | Cited, crawlable web content | Put real facts in crawlable text on your own site (Lever 4) |
How to do it: Claim every listing. Make name, address, phone, care types, and pricing posture identical across all of them — inconsistency is what makes an engine hedge. Fill in the fields operators skip: care levels offered, accepted payment types, amenities, hours. This is unglamorous and it’s the highest-leverage week of work you can do, because it feeds every other lever.
Lever 3 — A steady reviews motion
What it is: Reviews are both a ranking input the engines weigh and a trust signal families read in the AI’s own summary. Google reviews feed AI Overviews directly; Yelp feeds the ChatGPT side. A community with a handful of old reviews looks risky to both an algorithm and a daughter.
How to do it: Make asking for a review a routine step after positive family touchpoints — a successful move-in, a care-plan review that went well, a family event. Respond to every review, good or hard, in a way a prospective family would find reassuring. You’re not gaming a number; you’re maintaining a steady, recent flow of authentic signal on the exact properties (Google, Yelp) the engines read. A trickle that never stops beats a one-time push that goes stale in six months.
Lever 4 — A website with facts in crawlable text and live availability
What it is: Perplexity and the AI engines cite content they can actually read. If your differentiators live only inside a PDF brochure, a slow-loading image, or a “request info” wall, they’re invisible to the models — and to the families who never get the brochure.
How to do it: Put the facts families and engines need in plain, crawlable text on your pages: care types, floor-plan and pricing ranges, what’s included, neighborhood and transportation, staffing approach, and — critically — current availability. Availability is a direct-demand magnet: a family that learns you have an opening today calls today instead of filling out three aggregator forms. Answer the real questions families ask (“memory care vs. assisted living,” “what does the monthly fee include,” “can a couple with different care needs live together”) in text, not buried in a gated funnel.
See where you stand before you start. Run the free AI-visibility check → — it shows your per-engine visibility rate so you fix the engine that’s actually ignoring you, not the one you assume.
Lever 5 — Nurture past inquiries and professional referrals
What it is: The cheapest direct move-in is the family that already raised their hand. With a decision cycle that runs anywhere from 107 to 400 days, the prospect who said “not yet” eight months ago may be ready now — and if you stayed in touch, they come to you, not an aggregator. The same goes for professional referral sources: discharge planners, geriatric care managers, elder-law attorneys, and physicians who send families directly.
How to do it: Run a real nurture sequence against your past-inquiry list instead of letting it rot in the CRM — periodic, genuinely helpful touches, not just “are you ready yet.” Maintain the human relationships with local professionals who refer; those referrals arrive direct and pre-trusted. Both motions are fee-free by construction, and both are invisible in your reporting unless you’re tracking source (which is the next section).
Lever 6 — Be a citable source (GEO)
What it is: Generative Engine Optimization — earning your way into the sources AI engines cite when families ask. It’s the connective tissue across the levers above: the more you’re a clear, consistent, well-reviewed, fact-rich, referenced entity on the web, the more confidently engines name you. For the senior-living-specific version, see GEO for senior living.
How to do it: Beyond your own site and listings, become referenceable elsewhere — accurate presence in local and care-type directories, genuinely useful content that other pages and Reddit threads cite, and a consistent entity footprint so engines can connect the dots between mentions. You’re not chasing a backlink count; you’re making it easy for a model to conclude, from many corroborating sources, that you’re a real and relevant answer to a family’s question.
The part that actually matters: prove it drove the move-in
Here’s where most direct-channel efforts quietly die. You do the listings work, the reviews, the GEO — and then a budget review asks, “What did that get us?” If your answer is a shrug, the spend goes back to the aggregator line item that can show a move-in next to an invoice.
You cannot prove a direct channel works — or defend its budget — without closed-loop, directional attribution. Two things make this hard in senior living, and you have to design around both. First, the decision cycle runs 107 to 400 days: the AI-sourced inquiry in January becomes the move-in in October, long after anyone remembers the first touch. Second, “how did you hear about us?” is unreliable — families forget, conflate sources, or just say “the internet” when they actually started in ChatGPT. A self-reported field can’t carry a budget decision.
The fix is to join AI-sourced inquiries to move-ins through your CRM as directional, match-confidence attribution — not a false claim of perfect last-click certainty, but a defensible, confidence-rated link from first AI-sourced touch to move-in. That’s what lets you say “direct AI-sourced inquiries produced this many move-ins this period” with a number you’d stand behind. See senior living marketing attribution for the model and how to track move-ins by source for the mechanics.
Pair the two halves and you have a real program: measure each engine’s visibility rate with the free checker so you know where you’re being named, and use directional attribution to prove which of those named inquiries became fee-free residents. That’s how a direct channel earns — and keeps — its budget, one move-in at a time.