AI Readiness Score: where most independent practices actually sit in 2026

41 out of 100.

That is where the average independent healthcare practice sits when scored against the CARE Framework — the four-pillar assessment we use to evaluate how ready a practice is to implement AI across its workflows. It is not a failing grade in the catastrophic sense. The practice is functioning. Patients are being seen. Revenue is coming in. But it is below the threshold at which AI investment reliably delivers measurable returns rather than becoming another software subscription that gets paid for and not used.

The number comes from assessments across independent practices in physiotherapy, psychology, dentistry, occupational therapy, and speech and language therapy. The sample is weighted toward owner-operated practices of one to eight clinicians — the segment where AI is simultaneously most needed and most commonly over-promised. The 41 figure is a baseline, not a ceiling. The same practices scoring 41 today can be at 58 within 90 days without spending a dollar on AI technology. But to understand why that is possible, you need to understand what the score is actually measuring.

What the CARE Framework scores

The CARE Framework evaluates a practice across four pillars, each scored out of 25. The total gives the Practice AI Opportunity Score out of 100.

C — Clinical Workflow (0–25). How documented, standardised, and digitised are your clinical processes? A practice where every clinician uses their own note format, appointment types are inconsistent across the booking system, and treatment protocols exist only in the senior clinician's head scores low here. A practice with standardised note templates, clear appointment type definitions, and documented clinical pathways scores high.

A — Administrative Burden (0–25). How much administrative work is consuming clinician and staff time, and how clearly can you quantify it? This is not simply "how busy is admin" — it is whether you know which tasks take how long, who does them, and where the biggest time costs are. Practices with no time-tracking data, no workflow documentation, and no separation between clinical and administrative hours score low. The higher the documented admin burden, the higher the AI opportunity — which is why this pillar can score high even when the situation is difficult.

R — Readiness (0–25). Do your staff have the digital literacy to adopt new tools? Is your IT infrastructure capable of supporting cloud-based software? Do you have a data governance policy or privacy officer responsible for HIPAA compliance? Does the practice owner have both the appetite and the bandwidth to lead a change management process? Readiness is the pillar where most practices leak the most points, and the one that vendors consistently fail to assess before selling.

E — Economics (0–25). Does the practice have the financial capacity to invest in AI tools? Does it currently measure cost-per-workflow in any form? Can you identify how much a no-show costs, or how much clinician time costs per hour? Economic readiness is not about having a large budget — a solo OT with a tight budget who knows exactly what their admin is costing them scores higher on this pillar than a five-clinician dental group that has never calculated their no-show revenue loss.

Where the average score comes from: pillar by pillar

The 41/100 average breaks down unevenly across the four pillars. Administrative Burden typically scores highest — between 12 and 16 out of 25 — because most independent practices genuinely are carrying significant admin load and the evidence of it is visible during assessment. Clinicians are doing documentation until 8pm. Receptionists are chasing referrals and insurance authorizations for hours per day. The burden is real and it scores accordingly.

Clinical Workflow and Economics tend to score in the 9 to 13 range. Clinical workflows exist but are rarely documented in a form that could be handed to an AI system. Economics are felt but not measured — practice owners know no-shows are costing them money but very few have calculated the annual figure.

Readiness is consistently the lowest-scoring pillar, often landing between 6 and 10. This is the gap between a practice that is experiencing the problem and a practice that is positioned to solve it. Staff digital literacy varies considerably — typically one or two team members are confident with new software; the rest range from neutral to resistant. IT infrastructure is frequently a collection of systems that pre-date cloud-based working and were never designed to integrate. Data governance is almost universally informal: practices are HIPAA-aware but few have a written policy, a designated privacy officer, or a vendor vetting process for software that handles patient data.

The Readiness gap is why AI tools bought on the back of a compelling demo frequently fail in independent practice settings. The demo happens in a controlled environment with a vendor who knows how to use their own software. The real implementation happens in a practice where the receptionist has never used cloud software, the Wi-Fi drops three times a day, and the practice owner does not have three hours to lead onboarding. Readiness assessment before purchase is not optional. It is the variable that determines whether the investment works.

What a score of 41 means in practice

A score between 35 and 50 indicates what we classify as a practice in the preparation window. AI investment is premature if it is broad — buying multiple tools simultaneously, attempting to automate several workflows at once, or selecting tools based on feature lists rather than specific workflow problems. But targeted AI adoption in one carefully selected workflow is viable and can succeed.

The selection criterion at this score band should be simplicity of implementation. A clinical note AI that requires only a device with a microphone and an internet connection, produces a draft note for clinician review, and does not require PMS integration scores high on implementation simplicity. An AI scheduling system that requires integration with your PMS, configuration of booking rules, staff training, and patient-facing changes scores low. Both tools might deliver excellent outcomes in the right practice. At a readiness score of 41, only one of them is likely to stick.

The AMA's 2026 AI Adoption Survey found that physician practices reporting successful AI implementation were three times more likely to have started with a single use case than practices that attempted broad implementation from the outset. The practices that attempted multiple simultaneous implementations showed a 67% abandonment rate within 12 months. The data supports a sequenced approach, and it supports starting where implementation friction is lowest regardless of where the highest potential ROI sits.

So what for you: if your practice is in the 35 to 50 range, the first AI decision is not which tool to buy. It is which single workflow problem, if solved, would produce the most visible and measurable improvement. Pick that workflow. Find the lowest-friction tool that addresses it. Implement it properly. Then use the credibility of that success to fund and justify the next implementation.

What moves the score fastest

The readiness pillar is the most actionable in the short term, and improving it does not require AI spending. It requires operational investment — time, not money.

Documenting workflows is the single highest-leverage activity. Take the five administrative tasks that consume the most staff time and write down, step by step, how each one is currently done. This exercise routinely reveals inconsistencies between how different staff members perform the same task, which is itself a source of inefficiency. It also produces the documentation that any AI implementation will eventually require. A clinical note AI cannot be configured with custom templates if no standard templates exist. A scheduling AI cannot enforce booking rules if booking rules have never been written down. Workflow documentation is AI-readiness work and operational improvement work simultaneously.

Measuring administrative time is the second highest-leverage activity. A two-week time audit — where each staff member logs what they spend each 30-minute block of the working day doing — produces the data that makes an economic case for AI investment. It also shifts the conversation from "we feel like admin takes too long" to "admin tasks we cannot bill for consume 23% of clinician time and 68% of reception capacity." The second version of that sentence can justify a budget conversation. The first cannot.

Reviewing data governance is the third, and it is the prerequisite for any AI tool that handles patient data — which is almost all of them. A one-page data governance policy covering which software vendors have access to patient data, how BAAs are obtained and stored, and who is responsible for reviewing new software before purchase is sufficient for most independent practices. It takes two hours to write and eliminates the compliance exposure that stops many cautious practice owners from moving forward at all.

Collectively, these three activities — workflow documentation, time audit, data governance policy — typically produce a Readiness score improvement of 8 to 12 points within 60 days. Combined with the Economics improvement that comes from having actual time and cost data, a practice that enters this process at 41 frequently exits at 55 to 58, without having purchased a single AI tool. At 58, the same tools that would have failed at 41 have a reasonable probability of working.

The score that matters more than the average

The 41/100 figure matters as context. It tells you that the average independent practice is not AI-ready in any broad sense, and that vendors selling AI tools to independent practices are selling into a market where most buyers are not yet positioned to get value from what they are buying. That is a problem that benefits the vendors more than the practices — annual contracts get signed, tools get underused, and the vendor's churn rate becomes someone else's sunk cost.

The score that actually matters is yours. Two practices with identical headline scores can have completely different implementation profiles: one with strong clinical workflow documentation and a digitally confident team, the other with strong economics but poor readiness infrastructure. The same tool, at the same price, implemented at the same time, will produce different outcomes in each. The average tells you where the market is. Your score tells you where you are.

The CARE Framework assessment — conducted as part of the AI Opportunity and Growth Assessment — produces a scored breakdown across all four pillars, with specific gap analysis and a sequenced implementation roadmap. The point is not the number itself. The point is knowing which pillar is holding you back and what to do about it in the next 90 days.

A practice that enters an AI implementation at 41 and exits having spent $3,000 on tools that are not being used has not made progress. A practice that spends 60 days improving its readiness score to 58 and then selects one tool with precision has made a different kind of investment — one that compounds rather than depreciates.

The average is 41. The question is whether that describes where you are or where you used to be.

Related: Why 73% of AI tools go unused in healthcare practices · AI note tools in therapy: the compliance rules that now apply · The AI Opportunity and Growth Assessment