How AI Decides Which Businesses to Recommend
A Forest Acres family with a 6-year-old who can't read the chalkboard at school asks ChatGPT, "We need a pediatric-friendly optometrist in Forest Acres — preferably one that accepts our vision insurance and is patient with anxious kids. Any recommendations?" The AI returns two named practices with a short description of each. The third practice in the area — equally qualified, equally well-reviewed — is not mentioned.
Why? The AI weighed seven signals when it built that answer. Five of them favored the named practices. This article walks through all seven, using a Forest Acres optometry practice as the running example, with a self-audit you can use today to see exactly where you stand on each.
The Decision Set
7 signals
Across the four major AI assistants (ChatGPT, Perplexity, Google AI Overviews, Claude), the consistent pattern is seven signals weighed in roughly the same order — though weights shift slightly between platforms. Optimizing for all seven is the only path to consistent citation.
Signal 1: Entity Verification Confidence
Before recommending any business, the AI checks whether the business is a real, locally-operating entity. The signals it uses:
- Consistent NAP across 10+ surfaces (Google Business Profile, Bing Places, Apple Maps, Yelp, Healthgrades, ZocDoc, Vitals, BBB, the SC Board of Optometry registry, your insurance partners' provider directories)
- Verifiable business registration (SC Secretary of State filing accessible)
- Active Google Business Profile with recent posts, photos, and review activity
- Schema.org markup on the website that matches what other sources say
For the Forest Acres optometrist, the most common verification gap is an outdated provider directory entry from a previous practice location — the AI sees two addresses for what claims to be one practice and discounts confidence.
Common mistake: Letting old provider-network directory entries (VSP, EyeMed, Davis Vision, Spectera) stay outdated after a practice move or address change.
Time investment: 2-3 hours for a complete provider-directory + general-listing audit.
Signal 2: Topical Match to the Specific Question
Even after verification, the AI checks how well your business answers the specific question being asked. A general optometry practice is eligible for "optometrist in Forest Acres" but not necessarily for "pediatric-friendly optometrist in Forest Acres" unless your public content explicitly addresses pediatric care.
The signal: do your website, GBP attributes, services list, and review content contain the specific facets the question parses to?
For the example query, the practice needs to publicly state:
- "Pediatric eye exams" as a service category (on website + GBP)
- Insurance acceptance (which carriers, explicitly listed)
- Experience with anxious or first-time pediatric patients (in About page or reviews)
- Forest Acres as a stated service-area neighborhood (not just "Columbia metro")
Common mistake: Treating sub-specialties as marketing afterthoughts. If you handle pediatric, low-vision, dry-eye, or scleral-lens cases, each is a separate AI-citation opportunity that needs explicit content.
Time investment: 2-4 hours per sub-specialty to publish a proper service page.
The core principle: AI assistants reward specificity. The practice that explicitly states it handles pediatric, anxious patients, and Forest Acres families with a particular insurance is the practice that gets cited for that exact query.
Signal 3: Authority and Credential Signals
For healthcare and high-trust categories, credentials are explicitly weighed. The signals:
- State license verification (SC Board of Optometry registry, accessible)
- Specialty board certifications (American Board of Optometry, COVD for vision therapy)
- Professional association memberships (AOA, SCOA)
- Hospital affiliations if any
- Continuing-education evidence
- Named providers with credentials publicly stated
An optometry practice with "Dr. Sarah Chen, OD (SUNY College of Optometry, 2014), Board-Certified, member of the American Optometric Association and the South Carolina Optometric Association, fellow of the College of Optometrists in Vision Development" gets weighed differently than one with "Our team of caring eye doctors."
Common mistake: Burying credentials in PDFs or hiding them behind navigation. Credentials need to be on the About page and the provider-bio pages in extractable text.
Signal 4: Review Substance (Not Just Volume)
The shift from review-as-star-average to review-as-source-material is one of the biggest changes between 2022-era SEO and 2026 GEO. AI assistants now lift specific phrases from reviews into their answers.
For the Forest Acres optometrist, reviews that mention "Dr. Chen was patient with my anxious 6-year-old," "they had crayons and tablet games during the wait," "called us back same day with the prescription," and "Forest Acres location is easy to find with parking right out front" are the ones the AI extracts.
Reviews that say "great office!" — even at five stars — are summarized as generic positive sentiment, not extractable detail.
Common mistake: Asking for "a quick review" with no prompt for specifics. The patient writes the shortest acceptable thing.
Time investment: 5 minutes to rewrite the request template.
Signal 5: Cross-Surface Citation Density
How many independent sources mention you? The AI's confidence climbs when:
- Your business appears in 10+ directories with consistent data
- You have third-party mentions (local-news, community-event sponsorships, professional-association directories)
- Your social profiles (Facebook page, LinkedIn) are active and match your other data
- You appear in "best [category]" roundups by local-news or local-blogger sources
Citation density is what separates "the AI knows you exist" from "the AI confidently recommends you."
See Your Citation Density Right Now
Our free scan maps the cross-surface citation density of your business and benchmarks it against the top three competitors in your category.
Run Your Free Citation Density AuditSignal 6: Freshness and Activity
A business that has not posted to Google Business Profile in 14 months looks dormant. A website with stale dateModified values on schema markup looks neglected. AI assistants down-weight businesses that look inactive.
What "active" means:
- Google Business Profile post at least monthly (an update, photo, event, or offer)
- Service-page
dateModifiedvalues within the past 6-12 months - Recent reviews flowing (5-15 per month is healthy for an established practice)
- About-page provider list current (no providers who left a year ago still listed)
Common mistake: Setting up the website once and not touching it. Even small refreshes — updating a service page, adding a holiday-hours notice, posting a community-event photo — keep the freshness signal alive.
Time investment: 30 minutes per month is the minimum cadence to maintain freshness.
Signal 7: Bookability and Accessibility
The newest signal, still maturing in 2026: can a customer take action from the AI's answer? Practices that expose a public booking page, a clear "request appointment" form, or insurance-verification tooling are weighted higher because the AI can guide the user through a complete flow.
For an optometry practice, this means:
- A real online booking link (not a tel: link), preferably integrated with your practice-management system
- Public insurance-acceptance list with carriers explicitly named
- Stated new-patient policy (accepting / waitlist / next-available)
- Accessibility information (wheelchair access, low-vision-friendly intake, language support)
By 2027, agentic AI assistants will book appointments directly. Practices that have not exposed a bookable surface will be skipped.
Why Forest Acres practices have a window right now: The Forest Acres / Trenholm Plaza area is dense with optometry options, but most practices have made none of these seven moves. The practice that completes the seven-signal audit in the next 90 days will become the named recommendation for pediatric, family, and specialty queries — and stay there for 2-3 years before competitors catch up.
The Self-Audit (Run This Today)
Score yourself 0-2 on each signal (0 = not started, 1 = partial, 2 = strong).
- Entity verification: Is NAP consistent across 10+ surfaces?
- Topical match: Does each sub-specialty have its own service page with explicit facets?
- Authority: Are credentials in extractable text on the About page?
- Review substance: Do your last 20 reviews mention specific providers, services, and outcomes?
- Citation density: Do you appear in 10+ surfaces with consistent data, including 2+ third-party mentions?
- Freshness: Is your most recent GBP post within the past 30 days?
- Bookability: Can a new patient book online without calling?
Total score:
- 0-4: You are functionally invisible to AI. 60-120 days of foundation work will move you to the first citation.
- 5-9: You appear occasionally. Targeted fixes on the weakest signals can shift you to consistent citation in 90-120 days.
- 10-14: You appear regularly. Marginal improvements on the lowest-scoring signal compound the lead.
Get an Objective Score Across All Seven Signals
Our free scan rates your practice 0-2 on each of the seven signals, benchmarks against the top three Forest Acres competitors, and produces a ranked fix list.
Run Your Free Seven-Signal AuditThe Bottom Line
AI does not pick businesses to recommend at random. It weighs seven specific signals in the same rough order every time. The Forest Acres optometry practice that audits each signal honestly, then fixes the two weakest, will appear in citations within 60-120 days. The practice that ignores this for another year will keep wondering why a less-credentialed competitor keeps getting the AI's nod.
Start today: Run the self-audit. Whatever your lowest-scoring signal is, that is your highest-leverage fix.
Skip the Manual Audit
Our free scan runs the seven-signal audit automatically and emails you a prioritized 90-day plan.
Run Your Free Scan NowSources & Further Reading
- OpenAI: ChatGPT Search documentation (2024-2026)
- Perplexity AI: Source ranking and citation documentation
- Google Search Central: AI Overviews source-selection guidance (2024-2026)
- Schema.org: Optometric, MedicalBusiness, Person, Service, FAQPage type documentation
- American Optometric Association (AOA): Practice marketing and patient communication guidelines
- South Carolina Board of Examiners in Optometry: License verification registry
- South Carolina Optometric Association: Member directory
- VSP / EyeMed / Davis Vision: Provider-directory listing best practices
- BrightLocal: Local Consumer Review Survey (2024-2025)
- Heaston Innovations engagements: observed seven-signal outcomes across Midlands eye-care practices (2024-2026)
Note: The seven-signal framework is a Heaston Innovations synthesis based on observed AI behavior across engagements. Specific weights and behavior vary by platform and shift over time. The Forest Acres optometry example is illustrative.
Free Optimization Scan