The AI agent for the operating room.
Pare predicts what every surgery needs, what to open, hold sealed, and leave off, tuned to the surgeon and the procedure, from your own usage data. No patient data required. No rip-and-replace.
- No patient data
- Deterministic & explainable
- Audit-ready
- Human-overridable
- Orthopedic Drape Pack×1
- TKA Femoral Component×1
- TKA Tibial Baseplate×1
- Monocryl 3-0 Suture×3
- Silver Barrier Dressing×1
- Topical Thrombin Spray×1
- Barbed Closure Suture
- Bone Graft Substitute (5cc)
- Cement Restrictor Plug
Built with surgeons and perioperative teams at UVA Health
Grounded in the OR, not a generic template
Every recommendation combines surgeon behavior, procedure priors, and your institution's own utilization.
Surgeon behavior
Learns how each surgeon actually opens supplies, per procedure.
Patient contextRoadmap
The engine accepts case-specific clinical factors — BMI, ASA, EF, redo — with literature-validated rules. Activated once a data agreement covers clinical context.
Institution utilization
Grounds every recommendation in your own usage and catalog.
Your data stays yours
Start without sharing a single patient record
Pare's audit runs on utilization and schedule data alone — items opened, items used, surgeon, procedure. Nothing else leaves your building.
No patient data
No demographics, no diagnoses, no clinical characteristics. The audit never sees a patient record.
No BAA to start
The audit runs on de-identified utilization exports, so you can start without a business-associate agreement or a compliance review cycle.
No system integration
A one-time CSV export from your ASC platform. No interfaces, no IT project, no rip-and-replace.
Free audit first
See your waste number — recoverable dollars, per surgeon and procedure — before you commit to anything.
What Pare is
See the dollars, then let the agent act on them
Your cost data stops at the implant. Pare links every dollar to the case — the surgeon, the procedure, and what actually got used — in one place, not five.
01 · The free audit
See what your ORs are throwing away.
$1,294 in supply waste per case, industry-wide — most of it preventable. JAMA Surgery, 2025.
You already know what you pay per implant. Pare shows you what you pay per case — every dollar traced to the surgeon, the procedure, and the items opened but never used.
- Recoverable dollars, broken out per surgeon and per procedure
- Cost linked case by case — not just line-item PO totals
- One CSV export — no integration, no BAA to start
RoadmapPair utilization with your billing export to see per-case margin by payer, for keep-or-cut service-line calls.
- Surgeon opened this item in 88% of comparable cases
- Cross-surgeon procedure prior agrees across the service line
- Emergency safety floor enforced last (never Leave Off)
02 · The OR supply agent
The agent that rewrites the preference card.
Every night it reads tomorrow's schedule and stages a three-tier supply plan for each case — Open Now, Hold Sealed, Leave Off — tuned to the surgeon and the procedure. Your team reviews and overrides; the agent never acts alone.
- A plan per case, before the room opens
- Projected waste avoided, quantified in dollars per case
- Every tier explained, human-overridable, emergency floor enforced last
- Scanned tomorrow's OR schedule4 cases
- Generated supply plan — Total Knee ArthroplastyDr. Elena Vasquez
- Applied clinical safety floorspassed
- Staged for human reviewready
- Orthopedic Drape Pack
- TKA Femoral Component
- TKA Tibial Baseplate
- Monocryl 3-0 Suture
- Silver Barrier Dressing
- Topical Thrombin Spray
- Barbed Closure Suture
- Bone Graft Substitute (5cc)
- Cement Restrictor Plug
Same engine, different surgeon — real fixtures.
Built for the OR's highest bar
Explainable
Every tier ships with its reasoning.
Human-overridable
Your team has the final call, always.
Audit-ready
Every recommendation is logged and exportable.
Point of useRoadmap
Capture what's actually used, the moment it's used.
Today Pare predicts the plan and audits the waste. The next step: a two-second scan at the point of use reads each item's barcode, resolves it against the FDA GUDID reference, and reconciles what was actually used against what Pare predicted — so the prediction sharpens every case and what's used is captured for billing.
- Barcode-first — reads the label, never a patient record
- Resolved against the FDA GUDID reference, with a confidence score
- Predicted vs actual, reconciled per case — the loop closes
On the roadmap — not part of the pilot today. Supply data only; no patient records, ever.
How it works
How Pare works
A continuous loop — every case sharpens the next.
- 01
Learn
We ingest your center's case + preference-card export (HST Data Stream, SIS Data Feed) plus usage and outcomes to learn what actually happens.
- 02
Predict
Pare generates a tailored supply plan for your case the night before.
- 03
Execute
Your team reviews and executes with confidence in the OR.
- 04
Improve
Outcomes are captured and the model gets smarter, every time.
The problem
Why preference cards fall short
Preference card
One-size-fits-all.
Static lists assume every case is the same.
Pare
Case-specific.
Every prediction adapts to the surgeon, the procedure, and what actually happens.
Methodology
Our methodology
Pare uses a proprietary modeling approach combining machine learning with clinical expertise to predict case-specific supply needs.
Learn more about our methodologyFrom the operating room
There's an inherent fear among scrub techs of not having the appropriate supplies immediately ready and available.
The excessive supplies being left opened and unused at the end of each case in the OR should be addressed.
The preference card is the problem
Every OR runs on a card no one has updated in years. Pare rewrites it from what actually gets used — so the right supplies are open, and nothing else is.
Built with surgeons and perioperative teams at UVA Health
See what your ORs are throwing away.
See a case-specific supply plan built from your own data.