Be the name
AI gives.
Forty-five percent of consumers now ask an AI assistant to find a local business. It answers with two or three names. We measure whether yours is one of them, fix the specific reasons it isn’t, and make sure you answer the phone when it is.
No call. No signup. Results in about a minute.
Illustration of the pattern. Your real numbers are one scan away.
of consumers now use AI to find local businesses — up from 6% one year ago
of relevant local queries get a business named by ChatGPT at all
of calls to small businesses go unanswered — and most callers never call back
AI sends almost no traffic. It sends patients.
Two numbers get quoted constantly and they seem to contradict each other. Consumer use of AI to find local businesses jumped from 6% to 45% in a single year. Yet AI referral traffic is still under 1% of all website visits.
Both are true, and the space between them is the whole opportunity.
A search engine hands off a click. An AI assistant answers the question. The patient never visits your site — they read three names, pick one, and dial. The referral never appears in your analytics. The patient just says “I found you online.”
Every agency selling you on “AI referral traffic” is selling a 1% channel and will lose the ROI argument by month four. The channel that matters is the 45% one, and it is almost entirely uninstrumented. That is what we built an instrument for.
Find out where you actually stand.
We run real buyer questions against the sources AI engines ground on, check your listing footprint across the platforms they cross-reference, read your website the way a machine does, and score your eligibility against each engine’s observed rating floor.
We put three real buyer questions to ChatGPT with web search on and print the answers word for word, so you can read exactly what a patient reads. Every practice we name links to its source. The Benchmark runs sixty questions across Gemini, ChatGPT and Claude — which matters, because they disagree.
How an assistant decides who to name.
Not a ranking. A verification. The engine gathers mentions of you across sources, checks whether they agree with each other, and only commits to a recommendation when the corroboration holds.
Rating floors act as gates, not signals
ChatGPT has been observed recommending only businesses at roughly 4.3★ or above; Perplexity around 4.1★, Gemini around 3.9★. Google Maps will still surface you at 3.5★. Below an engine's floor you aren't ranked low — you're removed from the set.
Corroboration across sources you don't own
Engines assemble local recommendations from third-party directories, review platforms and forums — Yelp, BBB, Foursquare, niche vertical directories, increasingly Reddit and YouTube. Conflicting hours or service descriptions across them actively suppress confidence.
Your Google Business Profile is the spine
For proximity-first questions — "near me," "open now" — GBP is the primary currency across both Gemini and ChatGPT. Categories, service areas, attributes and hours are read as structured facts.
"Best in city" is a different competition
Proximity stops mattering and reputation takes over: review sentiment, inclusion in best-of lists, press mentions, topical depth. Most practices optimize only for "near me" and never enter this race at all.
Freshness is continuously re-scored
Recent posts, new photos, review responses, updated service descriptions. Roughly 85% of AI-recommended businesses turn over as models re-evaluate — visibility here is a position you hold, not a project you finish.
Each engine trusts different evidence
A source that carries real weight with Perplexity may barely register with ChatGPT. Treating "AI" as one channel is how practices end up strong on one engine and invisible on the others. We have measured a practice ranked #2 by one engine and named zero times out of fifty-eight by another.
We don’t sell AEO or GEO. Nobody should.
On May 15, 2026, Google published documentation that rejected “answer engine optimization” and “generative engine optimization” as disciplines separate from search, and named the tactics agencies had been packaging as specialties:
“From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO… many suggested ‘hacks’ aren’t effective. You can ignore tactics like ‘chunking’ content, creating unnecessary AI text files, or pursuing inauthentic mentions.”— Google Search Central, AI Optimization Guide
If someone is invoicing you for a GEO line item on top of an SEO retainer, that is not innovation. It is a markup.
What is real is narrower and harder: whether an assistant, asked a buyer’s question, says your name. That is measurable. It moves. And roughly 85% of it turns over as engines re-evaluate — which is exactly why a one-time rebuild decays, and why we sell an instrument and a program instead of a package of pages.
Three parts. One number.
The instrument
You cannot manage what nobody is counting. We run a standing panel of real buyer prompts across Gemini, ChatGPT and Claude, every week.
- 60+ buyer prompts, locked at onboarding
- Competitor set derived from your real market
- Per-engine breakdown, tracked over time
- Published methodology — you can check our work
The corrections
The work that actually shifts the number, in the order that matters. No chunking, no llms.txt, no bought mentions.
- Rating repair to clear engine eligibility floors
- Entity and NAP corroboration across 15 AI-cited sources
- Google Business Profile depth and weekly freshness
- Structured data for practice, providers and each service
- Question-first service pages that answer what's asked
The proof
Demand you generate and don't answer is demand you paid for twice. This is where the ROI becomes arithmetic.
- Every call, form and DM answered inside 60 seconds, 24/7
- Missed-call text-back with a booking link
- Structured follow-up on every unconverted inquiry
- Dormant patient reactivation
- Booked-appointment dashboard with source attribution
Recommendation Rate
The share of real buyer questions in which an engine names you. Tracked weekly across Gemini, ChatGPT and Claude, against the competitor set we derive from your actual market — not one you guessed at. Then tied to the thing you actually bank: booked appointments.
Published, because most aren’t.
Presence
The visibility engine. For practices already answering their phone well.
- Weekly Recommendation Rate tracking
- Rating repair system
- Entity sync across 15 AI-cited sources
- GBP depth + weekly freshness
- 6 service pages / month
Growth
Visibility plus Capture. Break-even is roughly two extra patients a month.
- Everything in Presence
- 60-second response on every inquiry, 24/7
- Missed-call text-back + booking
- Dormant patient reactivation
- 12 service pages / month
- Booked-appointment dashboard
Authority
Category exclusivity in your metro, contractually. One practice per category.
- Everything in Growth
- Exclusivity in your city + category
- 2 earned media placements / month
- Provider thought leadership
- Booking-flow conversion audit
- Up to 3 locations
Every program starts with a $500 Benchmark, credited in full to your first month.
If your Recommendation Rate hasn’t moved in 90 days, we work free until it does.
We publish the methodology and lock the prompt set at onboarding, so the number we’re graded on isn’t one we can quietly move. It also means we turn down practices we can’t help — under 4.0★, that’s a different job, and we’ll tell you so.