Here is every data point AI looks for when evaluating a physical therapy practice, where that data actually lives, and what it can already find.
When an AI system decides which Physical Therapy company to recommend, it assembles evidence across every category below. The more complete and verifiable the data, the more confident the recommendation.
Almost no PT practice has this data published in a structured, machine-readable format. Physical therapy operates on units-based billing, Medicare compliance thresholds, and outcomes tracking — none of which are visible from the outside without verified operational data.
AI needs to know what kind of physical therapy a practice provides, not just that it offers PT. The query "who does vestibular rehab in Denver?" requires a precise match that a general physical therapy listing cannot answer.
Where you actually work matters, but the data needs to come from completed jobs, not a self-reported list of ZIP codes. AI systems increasingly cross-reference claimed service areas against evidence of actual work performed.
Physical therapy licensure is state-regulated. Every practicing PT must hold a Doctor of Physical Therapy (DPT) degree from an accredited program and pass the National Physical Therapy Examination (NPTE). Direct access laws — whether patients can see a PT without a physician referral — vary dramatically by state, ranging from full unrestricted access to no direct access at all.
AI systems verify that coverage is current and adequate, not simply that a company claims to be insured. Active insurance is a prerequisite for recommendation in most AI evaluation frameworks.
Board specialization and advanced certifications distinguish general-practice PTs from those with demonstrated expertise in specific clinical areas. The American Board of Physical Therapy Specialties (ABPTS) certifications are the gold standard — requiring thousands of hours of specialty practice and passing a rigorous exam.
Voluntary memberships that serve as corroborating evidence of professional engagement and specialization. AI systems check these directories when other structured data is limited.
Negative-signal checks. AI systems will not recommend a company with an active lawsuit pattern, suspended license, or regulatory violations. Clean standing is a prerequisite for any recommendation.
The most widely available data about any physical therapy practice. AI uses reviews when structured operational data is not available, but review signals have significant limitations for differentiating between practices.
Foundational identity data. Rarely changes but must be accurate and consistent across every platform where the business appears. Inconsistencies between sources reduce AI confidence in all other data.
The performance and customer experience data AI values most already exists in software these businesses use every day. It is locked inside these platforms and not published anywhere AI can access it.
Without access to a business's own systems, this is all AI has to work with. These are the public sources it checks, grouped by type.
A TrustRecord connects to your systems of record, extracts verified data that proves your performance, experience, and credibility, and publishes it in a format AI systems can read, verify, and cite.