P3 Health Partners (PIII): AI Point-of-Care Tool Closes Nearly 90% of Care Gaps Where Used
Summary
P3 Health Partners says clinicians address nearly 90% of care gaps where its AI point-of-care tool is used, but it is open in only roughly half of eligible visits.
P3 Health Partners (PIII) said on its August 10, 2026 earnings call that clinicians using its AI point-of-care tool addressed nearly 90% of care gaps during the visit itself. The tool was open in only roughly half of eligible visits, and the company tied it to no cost figure[1].
What P3 does, and what the tool does
P3 Health Partners (PIII) is a physician-led population health company. It contracts with Medicare Advantage plans, takes a fixed per-member premium, and pays for the medical care its attributed patients actually use; the difference is where its profit comes from. Revenue depends on how accurately a patient's condition is documented. Cost depends on whether chronic disease is kept under control.
The AI point-of-care tool aims at both. While the patient is still in the exam room, it pushes that patient's risk predictions, diagnosis codes that have not yet been fully captured, and the checkups, screenings and chronic-disease follow-ups that are due but have not been done — the care gaps — in front of the clinician, prompting the clinician to handle them on the spot. The users are P3's contracted clinicians and provider partners, the tool sits inside the interface they already work in every day, and it occupies a core operating position.
From a 2024 partnership to a first measurable readout
The public starting point is the May 8, 2024 earnings call. P3 announced a strategic partnership with health data company Innovaccer to use its AI platform for predictive risk modeling and to feed prompts into the clinician workflow, with InNote, an EHR-agnostic plug-in, as the provider-facing entry point. Management set two boundaries at the same time: full implementation would take 12 to 18 months, and the arrangement would be "cost neutral"[2] — a statement about what P3 would spend, not about what it would earn back.
By the August 10, 2026 call, the same workflow carried a measurable readout for the first time, and it was rolling out faster than originally planned[1]. The Innovaccer and InNote names no longer appear in the public description, but the users, the point in the visit where the tool intervenes, the information it handles and the prompts it produces are unchanged, and the workflow that surfaces care opportunities and quality gaps to clinicians has also been described consistently in the company's annual filings. This is the same application maturing.
The numbers measure different things
Nearly 90% is a conditional figure. It holds only in the visits where the tool is open, and the company did not say what the other half achieves. Paired with it is a documentation capture rate running several points ahead of the company-wide average, a comparison that points the difference back to the tool itself[1].
More than 65,000 lives is the number of members the tool reaches, not a count of uses. The roughly half open rate cuts the applicable scope of the first result in half; P3 attributes that to how the tool was rolled out and is replacing remote onboarding with in-office training staff placed directly in provider offices[1].
The path to the financials is drawn, not measured
The route from this operating improvement to the financial statements is not hard to sketch. More complete documentation brings risk-adjusted per-member payment closer to a patient's true condition. Care gaps closed during the visit keep chronic disease and missed screenings from escalating into emergency visits or admissions. Both ends land on medical margin — premium revenue less medical expense — and on the medical cost ratio behind it.
For now that is only commercial logic. Nearly 90% is a workflow output, not a financial result. P3 has not disclosed the difference in medical spend or risk-adjusted revenue between the patient groups that used the tool and those that did not, and it has not separated the tool's contribution out of the medical cost ratio.
What is confirmed, and what would settle it
What can be confirmed today is that for a company carrying medical risk on a per-member basis, this tool compresses documenting a condition and closing a care gap into a single action, and in the visits where it is used the effect is clear. What it is ultimately worth depends on how often clinicians open it, and by the company's own measure that is currently happening about half the time.
Turning the improvement in the workflow into an improvement in the financials requires P3 to give a concrete difference in medical spend or risk-adjusted per-member revenue between the using and non-using groups, or to quantify the tool's effect on the medical cost ratio once the open rate rises.
Application assessment
- AI Point-of-Care Population Health Tools | Business role: core operations | Deployment stage: limited production | Scope: multi-business or multi-region | Value type: cost reduction
Sources
[1] Drillr · P3 Health Partners Inc. (PIII) · 2026-08-10 · Earnings call
"Where the tool is in use, the results are clear. Providers are addressing nearly 90% of care gaps at the point of care, with capture rates running several points ahead of our broader enterprise average."
[2] Drillr · P3 Health Partners Inc. (PIII) · 2024-05-08 · Earnings call
"And it's going to take 12 to 18 months to implement the full partnership to be in effect, and it will be cost neutral for us as well."