What AI Looks for Before Recommending a Business

Updated May 2026 • 9 min read

A Lexington adult with worsening dry-eye symptoms after years of contact-lens wear opens ChatGPT on a Thursday evening and asks, "I'm in Lexington SC and have moderate dry eye that's gotten worse the past few months — I need an independent optometrist who actually treats dry eye (LipiFlow, IPL, scleral lens fitting if needed), takes BCBS Vision, and isn't a big-box chain." Two optometrists appear in the answer. Several others in the corridor that could have helped are not mentioned — even some that genuinely treat dry eye. The difference comes down to specific verifications the AI runs before placing a business in a recommendation.

"What AI looks for" is not a single criterion — it's a verification process. This article walks through the checks AI assistants typically run, in roughly the order they happen.

The Verification Sequence

When an AI assistant processes "independent optometrist in Lexington who treats dry eye," it runs through approximately this sequence:

Verification 1: Does the business exist?

Confirmation that the business is real, operating, and locatable. Sources checked:

• Google Business Profile (claimed, verified, current).
• State licensure registry (SC Board of Examiners in Optometry).
• Bing Places, Apple Maps, Yelp.
• Insurance-network provider directories (VSP, EyeMed, BCBS Vision).

If the business does not appear on the major Tier 1 platforms or is unclaimed on key directories, the AI may hedge or skip it.

Verification 2: Is the business in the right category?

Primary category alignment to the query. For "optometrist," the AI checks:

• GBP primary category = Optometrist (or Eye Care Center).
• Schema.org type on the website (Optometric, MedicalBusiness with relevant medicalSpecialty).
• Insurance-network classification.
• State-licensure category (Optometrist vs Ophthalmologist).

Verification 3: Is the business in the right location?

Geographic match. The AI checks:

• Operating address in Lexington (or adjacent service area).
• Service area declarations matching the query.
• NAP consistency confirming the location is current.
• Recent reviews from local customers confirming actual operation.

Verification 4: Does the business actually do what the query asks?

Specialty match. For "treats dry eye (LipiFlow, IPL, scleral lens fitting if needed)":

• Service-page content explicitly naming the relevant treatments.
• Schema declarations of those services.
• Reviews mentioning the same treatments.
• Provider bio mentioning relevant training or specialty experience.

This is where most non-recommended businesses lose the citation: they "treat dry eye" in their actual practice but their published content does not specifically establish that they offer LipiFlow, IPL, or scleral lens fitting.

Verification 5: Does the business accept the query's insurance/payment?

For "takes BCBS Vision":

• Insurance-accepted list on the practice website.
• BCBS provider-directory listing (cross-verified).
• VSP, EyeMed, or other vision-plan directory matches.

Verification 6: Does the business match the query's "type" qualifier?

For "isn't a big-box chain":

• Business classification (single-location independent vs. chain location of LensCrafters, Visionworks, etc.).
• Brand and identity signals.
• Ownership language on the about page.

Verification 7: Is the information consistent and trustworthy?

• NAP consistency across all platforms checked.
• Credential claims that cross-check externally.
• Review patterns that look authentic.
• No major flagging from BBB, regulatory complaints, etc.

Verification 8: Is the business recently active?

• Recent GBP posts (within last 90 days).
• Recent reviews flowing in (3-8 per month is healthy).
• Owner-responses to recent reviews.
• Updated date stamps on key pages.

Verification 9: Are there third-party trust signals?

• AAO (American Academy of Optometry) or AOA (American Optometric Association) membership.
• Specialty designations (FAAO, ABO, COVD if applicable).
• Local-news mentions or community involvement.
• BBB accreditation or favorable rating.

Verification 10: Is the named-author/provider clearly established?

• Named optometrists with bios, credentials linked, schema markup.
• License numbers cross-verifiable via SC Board of Examiners in Optometry.

Verifications 11-12: Topical depth and authority indicators

Beyond the basic checks, the AI evaluates whether the business has demonstrated topical depth on the specific specialty (dry eye, scleral lens fitting) and whether external authority signals exist (trade-press mentions, named-doctor publications, community recognition).

Where Most Local Businesses Fail the Verification Checks

Common failure points, in approximate order of frequency:

Failure 1: Specialty match (Verification 4)

The most-common failure mode. A practice that genuinely offers dry-eye treatment has either no service page about it, or a single thin paragraph mentioning it, or services listed only in a footer or sidebar without dedicated content. The AI cannot verify the specialty match, so the business gets hedged or skipped.

Fix: Dedicated service pages for each specialty offering, 1,200+ words, with named treatments, named-provider association, schema markup, and FAQ.

Failure 2: Information consistency (Verification 7)

The second most-common failure. NAP inconsistencies across platforms, unclaimed credentials, or contradictions between what the practice says about itself across different surfaces.

Fix: Quarterly NAP audit. Pick canonical versions of every business-identity field; enforce consistency across all 12-18 platforms.

Failure 3: Recent activity (Verification 8)

A practice that completed setup in 2023 and hasn't updated GBP since often fails the activity verification. Stale-business signals trigger hedging.

Fix: Weekly GBP posts. Monthly content updates. Continuous review-pipeline cadence.

Failure 4: Named-provider establishment (Verification 10)

Practices with "Our Team" content and no named-provider bios with verifiable credentials lose Verification 10. Even when individual optometrists are credentialed, the AI cannot extract their identity from generic team-page content.

Fix: Each provider has a dedicated bio page with credentials, license-verification link, Person schema with hasCredential, and association to specific services in body content.

Failure 5: Third-party trust signals (Verification 9)

Practices that exist online but have no AAO/AOA membership visibility, no BBB record, no community-involvement signals fail to provide external verification of standing.

Fix: Pursue and display industry memberships and credentials. Maintain BBB profile. Build at least one community-presence signal per quarter.

How To Pass All the Verifications (Practical Build for an Optometrist)

For a Lexington independent optometrist who wants to be the AI's default recommendation for dry-eye queries:

Months 1-2: Foundation Verifications (1, 2, 3, 7)

• Complete and verify GBP with primary category "Optometrist" and secondary "Eye Care Center."
• NAP audit across Tier 1 platforms (GBP, Bing, Apple Maps, Yelp, Healthgrades, Vitals, Zocdoc).
• Confirm SC Board of Examiners in Optometry registry entry is current.
• Confirm insurance-network provider listings (VSP, EyeMed, BCBS Vision) are accurate.

Months 3-4: Specialty Verifications (4, 5)

• Build a dedicated dry-eye service page (1,500-2,000 words) with named treatments (LipiFlow, IPL, scleral lens fitting, meibography, omega-3 protocols, etc.), named-provider association, Service schema, and FAQ schema.
• Add additional specialty pages for myopia management, contact-lens fitting, and other distinct services.
• Insurance-accepted page listing each plan with what's typically covered for vision exam and medical-eye care.

Months 5-6: Provider Verifications (10) + Activity (8)

• Build dedicated bio pages for each optometrist with credentials, SC license number linked to verification, AAO/AOA memberships, specialty fellowships, Person schema with hasCredential.
• Each bio links to the services that provider performs; each service page links to the relevant providers.
• Establish weekly GBP posting cadence.
• Rewrite post-visit review-request template for specificity coaching.

Months 7-9: Trust + Authority (9, 11-12)

• Confirm AAO/AOA membership displays.
• Pursue BBB accreditation.
• One sponsorship or community-event participation with web coverage.
• Pitch one trade-press or local-news commentary opportunity.
• Re-run the four-assistant prompt test for dry-eye and other priority queries.

By month 9-12: the practice passes all 12 verifications cleanly. AI citation for dry-eye, scleral lens, and related specialty queries becomes consistent.

What "Almost Passes" Looks Like

Many practices "almost pass" — meaning they're in the AI's candidate pool but don't get named in the final answer. Common almost-pass profiles:

Profile 1: Strong foundation, weak specialty content

The practice has GBP, NAP consistency, current insurance listings, but service pages are generic ("we offer comprehensive eye care"). The AI verifies existence and location but cannot verify specialty match. Gets retrieved but rarely cited for specialty queries.

Profile 2: Strong content, weak credentials/authority

The practice has substantive service-page content but providers are anonymous ("our experienced team"). The AI extracts the content but can't anchor it to credentialed humans. Gets hedged language in the AI's mention.

Profile 3: Strong credentials, weak recency

The practice has excellent credentials and historical authority, but GBP hasn't been touched in 18 months, last review was 8 months ago, content shows 2023 dates. The AI flags it as "possibly inactive" and prefers more current alternatives.

Profile 4: Strong on-site, weak external verification

The practice's website is well-built, but external sources (Healthgrades, Vitals, BCBS provider directory, AAO directory) either don't list the practice or list outdated information. The AI's cross-verification finds inconsistency and hedges.

The Bottom Line

AI assistants run a verification process — not a single check — before recommending a local business. The Lexington optometrist who passes all 8-12 verifications cleanly gets named when the patient with worsening dry eye asks ChatGPT on a Thursday evening. The practice that passes 7 of 12 and fails on specialty match, provider establishment, or recent activity does not — and the AI's pass/fail approach to verification is what most owners haven't fully internalized when diagnosing their visibility.