This is a Path B operator case-study piece, with the dental space leaning real-case-heavy because Pearl and Overjet both publish substantial small-and-family-practice customer rosters with named outcomes. Every vendor-published outcome is cited as such and not as independently audited measurement.

If you run a 1-to-5 dentist practice and you want to know what your peers have actually deployed in 2026, the published corpus across the two FDA-cleared dental AI platforms (Pearl and Overjet) plus the dental practice management vendors (Dentrix, Eaglesoft, Planet DDS, Carestream) is now substantial enough to read named cases from. Three patterns are clear: AI radiography assist now ships at solo-practice scale with FDA clearance, revenue-cycle automation (insurance verification, billing) is producing the biggest published time savings, and ambient voice AI for clinical notes is the next surface to ship widely.

The published corpus, with named cases

Pearl is “the only FDA-cleared AI tools that support both 2D X-rays and 3D CBCT imagery” per its own published page, with 26+ regulatory clearances across multiple countries and “more than 23,000 practices” in its customer base. Pearl’s product surface includes Second Opinion (pathology detection on 2D X-rays), Second Opinion 3D (CBCT review), Practice Intelligence (data-to-growth analytics), Precheck (cost estimates and coverage), Calibrate (AI-guided clinical training), and Pearl Voice (ambient voice AI for clinical notes). Pearl’s published case studies include Rand Center for Dentistry, Onsite Dental, and Blue Court Dental, though the published material does not always specify practice size.

Overjet publishes a substantially larger named small-and-family-practice roster, including Midtown Dental Studio, Kids First Pediatric Dentistry, Saddler Family Dentistry, Frankel Dentistry, Westport Family Dental, Smiles4Kids, Dental Care of Michigan, Scottsdale Smile Center, Prentice Dental, Great Smiles Family Dentistry, The Airway Dentists, New England Family Dentistry, Baines Family Dental, Leading Dental Solutions, Heber City Dental, Perio Atlanta, Total Dental Care, Promenade Center for Dentistry, Quest Dental, Oceanic Dental, and Cape Dental Care. Overjet’s product surface covers caries detection (Vision AI, FDA-cleared), patient-education visualization, ambient clinical documentation, insurance verification automation, and revenue cycle management.

Specific published outcomes from Overjet’s customer page include:

Promenade Center for Dentistry reports saving 20 hours per week on insurance verification.
Quest Dental reports increased Crown production by 19% and restorations by 15%.
Affinity Dental Management reports Crown treatment planning rising from 11% to 26% acceptance.
Midtown Dental Studio reports case acceptance improvement of 566%.

These are vendor-published claims, not independently audited measurements. The 566% case-acceptance figure in particular is the kind of suspicious-precision number a careful reader should treat as marketing rather than baseline-controlled measurement. The directionality is real (AI-assisted patient education does materially improve case acceptance in the published corpus); the specific percentage is a single data point.

The pattern across the corpus: where AI ships at small-practice scale

Reading across both vendors’ published cases, four workflows now show consistent solo-and-family-practice AI deployment:

1. AI-assisted radiography review. Both Pearl Second Opinion and Overjet Vision AI are FDA-cleared for caries (cavity) detection on 2D bitewing X-rays. The pattern is: dentist takes the X-ray as normal, AI overlays detected pathologies, dentist confirms or dismisses each finding before discussing with patient. The clinical claim is “second pair of eyes that doesn’t get tired”; the operational claim is “more confident treatment recommendations because the AI confirmed what you saw.”

2. Insurance verification automation. This is where the largest published time-saving figures sit. Promenade Center’s published 20 hours per week on Overjet’s insurance verification feature is the single biggest single-workflow time claim in the dental-AI published corpus in 2026. The work being automated is the back-office staff time spent calling insurers to verify coverage and benefits before treatment, which historically takes 15-30 minutes per patient and now takes minutes via AI-driven payer-portal automation.

3. Patient-education visualisation for treatment acceptance. Both Pearl and Overjet ship “show the patient what the AI sees on the X-ray.” The published case acceptance lifts (e.g. Affinity 11% → 26% on Crowns) suggest this is the lever that converts AI-detected pathology into accepted treatment plans. This is also where the small-practice ROI argument actually closes: a solo practice that adds one accepted Crown per week from improved case-acceptance pays for the AI subscription many times over.

4. Ambient voice AI for clinical notes. Pearl Voice and Overjet’s clinical documentation are both shipping in 2026; the Overjet customer roster does not yet attach specific outcome figures to this workflow but the Pearl product page positions it as a peer to the radiography surface. The published bookkeeping-style time-savings claims are not yet in the corpus for ambient voice in dental.

What the corpus says does not yet ship reliably at small-practice scale

The published corpus is also clear about the line.

AI as the diagnostic decision-maker. Both Pearl and Overjet are FDA-cleared as decision-support, not as autonomous diagnosis. The dentist still makes the call; the AI helps confirm it. The marketing language is careful about this; the published cases do not claim AI is replacing diagnostic judgement.

Treatment planning beyond what the AI was trained for. AI is trained to detect specific pathologies (caries, bone loss, periodontal disease patterns); it is not trained to recommend orthodontic plans, prosthodontic options, or complex multi-stage treatment sequences. The published cases stay carefully within scope.

HIPAA-defensible AI for non-FDA-cleared workflows. Solo-practice dentists experimenting with consumer ChatGPT Plus or Claude Pro for treatment-letter drafting or patient-communication summarisation are entering the same HIPAA-risk territory as any other US healthcare provider using consumer AI. The published platforms (Pearl, Overjet) are HIPAA-compliant by design with signed BAAs; the consumer AI tier is not, regardless of how thoughtful the dentist is about prompts.

Reading 3D CBCT scans at small-practice scale. Pearl Second Opinion 3D ships, but CBCT machines themselves are not standard at solo-practice scale (they typically live at oral surgery, endodontic, or larger group practices). The published 3D AI cases skew toward those settings, not toward the 1-2 dentist family practice.

The 2-dentist family practice: defensible 2026 stack

For a 1-to-3 dentist practice asking “what should we run in 2026,” the published corpus suggests a defensible stack of two AI platforms plus the practice management software you are already on:

One FDA-cleared radiography AI (Pearl Second Opinion or Overjet Vision AI). Both are competitive on caries detection on 2D bitewings; pick on integration with your existing imaging software (Dentrix, Eaglesoft, Carestream, Planet DDS). The named small-and-family-practice rosters on both vendor sites confirm this is now solo-and-family-practice-scale tooling, not enterprise-only.
One revenue-cycle AI feature, either bundled into the radiography vendor (Overjet’s insurance verification + RCM) or via your PMS vendor’s own AI. The Promenade Center 20-hours-per-week claim is the published high-water mark; honest conservative reading is “meaningful staff time saving on benefit verification.”
An Enterprise-tier general AI assistant (Anthropic Enterprise or equivalent) for non-clinical-PHI work like marketing copy, response to non-PHI patient-portal questions, or internal staff training material. Not for any workflow where patient PHI is in the prompt.

What the corpus says you should not run for clinical work: consumer Claude Pro or ChatGPT Plus for anything touching PHI. Free-tier any-vendor AI for clinical notes. Non-FDA-cleared “AI radiography review” tools that have not gone through the De Novo or 510(k) process the FDA requires for diagnostic decision support.

What we are deliberately not claiming

We are not claiming any specific named practice from the Overjet or Pearl roster will produce the same outcomes if implementation is replicated. The published cases are vendor-curated and skew toward early-adopter practices with the operational maturity to deploy AI well; a practice without that maturity will see different (likely smaller) outcomes.

We are not claiming AI replaces a dentist, dental hygienist, or dental assistant. The four workflows above accelerate work that those roles already do; the dentist’s diagnostic and treatment-planning judgement remains the practice’s value.

We are not claiming the specific Overjet outcome figures (20 hours/week, +19%, 566% case acceptance) are typical or generalise outside the named practice. These are vendor-published case-study figures and should be read as evidence that the workflow exists and produces meaningful outcomes for some practices, not as a benchmark every practice should expect.

What changes this read

Cadence on this piece is 60 days because the dental-AI vendor surface is shipping fast (especially in voice AI and 3D imaging) and the FDA-clearance landscape evolves as new modalities clear. The three things that would change the verdict:

An independently audited (non-vendor-published) measurement of small-practice AI outcomes lands. Currently the published outcomes are all vendor-curated; an independent measurement (from the ADA, an academic dental school, or a third-party study) would let us calibrate the vendor figures.
A new FDA clearance materially expands the radiography AI surface. Currently caries detection and bone-level analysis are the cleared workflows; clearance for a new modality (e.g. AI-assisted ortho diagnosis) would expand the small-practice deployable workflow set.
A consumer-tier AI vendor ships HIPAA-eligible inference. Currently the line between “Enterprise tier with BAA” and “consumer tier” is the determining factor for what dental practices can use AI on. A consumer tier with HIPAA eligibility (Anthropic Enterprise has a HIPAA-ready option; OpenAI’s Enterprise tier is similar) becoming standard at consumer pricing would change the small-practice ROI math.

We will re-test against the Overjet customer page and the Pearl product pages on or before 26 Jun 2026.