It doesn’t wait to be asked. It drives the visit.
ChironAI™ is an Agentic Clinical Operating System, and Chiron is the Digital Employee who runs it. Chiron works an encounter the way a strong resident does — it finds the patient, reads the whole chart, pulls the labs that matter, checks safety, reasons the differential, and drafts the assessment and plan — taking initiative through a loop of tool-calls you can watch and reasoning you can read. Everything it produces is a draft; the physician reviews and attests, and nothing is final until it’s signed. Powered by Eve-Healthcare™ F5/reasoner.
Not a spinner. You watch Chiron work.
Chiron doesn’t wait to be prompted. It drives the encounter as a plan — finding the patient, reading the whole chart, pulling the labs that matter, checking safety — and it shows every tool it calls and every step of its reasoning as it goes. What it produces is a draft. You review it, and you attest. Nothing enters the chart until you sign.
- 01Locate
Chiron opens the encounter
- 02Gather
It reads the whole chart, not the last note
- 03Retrieve labs
It pulls the labs that matter, and flags what’s off
- 04Check safety
It checks the interaction before it proposes the order
- 05Reason & draft
It drafts the assessment, plan, and orders — for your signature
Illustrative — representative of product UI. Synthetic case data; not from any real patient. Tool-calls and live reasoning reflect the real product posture.
Every surface runs the same loop. The physician closes it.
Whether it’s an intake, a lab panel, or a radiology read, Chiron runs the same command-first loop — and the loop always terminates at a human decision. That is the boundary between agency and autonomy the architecture is built to hold.
- 01
Chiron takes the encounter
From a command or a schedule slot, Chiron opens the workspace and sets a plan for the visit. The encounter is command-first and Chiron-driven, not a form to fill.
- 02
It calls tools you can watch
find_patient, read_record, search_labs, drug_interactions — Chiron reaches for the record, the labs, and the safety checks itself. Every tool-call is visible, so you see exactly what it looked at.
- 03
It reasons in the open
Live reasoning streams on every AI surface: the differential, the discriminating features, the safety checks it runs, and the blind spots Chiron names for itself. No black box, no verdict from nowhere.
- 04
It drafts the artifact
Assessment, plan, orders, the lab read, the radiology impression, the regulatory form — Chiron assembles a structured draft, with every clause traceable to the record field that produced it.
- 05
You review and attest
Nothing enters the chart on Chiron’s say-so. Each artifact carries a must-review-before-final gate; the licensed clinician reviews, edits, and signs. The copilot asks approval before it acts.
- 06
The loop informs the next step
What you attest updates the Chiron-maintained patient snapshot and the curated dashboard, so the next encounter starts from a current, reasoned picture — not a stale chart.
Decision support, not diagnosis. Chiron is not a licensed clinician and does not practise medicine; it surfaces, reasons, and drafts, and every clinical artifact carries a must-review-before-final gate. The licensed clinician decides — per AB 3030 / SB 1120 disclosure and AB 489.
What the product actually does, in eight deep-dives.
The loop above runs everywhere; these are the eight surfaces where it does the most work. Each entry is anchored to the actual platform implementation — the reasoner’s disposition, the consultant routing, the workflow rule the system enforces. If a CMIO asks “how does it work?”, this is the answer.
- 01Reasoning
Abductive clinical reasoning
Inference to the best explanation — the cognitive mode clinicians actually use to construct a differential diagnosis. Eve-Genesis (Clinical Edition) trains the reasoner in it specifically.
Read deeper → - 02Vision
The vision route for radiology
A dedicated vision model routes image-bearing requests through modality-appropriate frameworks (BI-RADS, LI-RADS, PI-RADS, Lung-RADS, others). Structured reports with provenance.
Read deeper → - 03Documentation
Source-grounded documentation
SOAP notes assembled from the encounter’s structured evidence — not generated as opaque prose. Editable, traceable, attestable.
Read deeper → - 04Boundaries
Decision-support boundaries
ChironAI is decision support, not decision-making. The boundaries are architectural — there is no surface in the platform that issues a treatment plan.
Read deeper → - 05Pharmacology
Pharmacology and drug interactions
Full medication-schema reasoning. Indication, mechanism, contraindication, drug-drug interaction (4-tier severity), drug-disease interaction, dose-adjustment guidance.
Read deeper → - 06Risk
Calibrated risk stratification
Wells, GRACE, MELD, CHA₂DS₂-VASc, and other frameworks with the explicit reasoning that produces each score.
Read deeper → - 07OM workflow
Occupational medicine regulatory workflow
AOE/COE causation, DWC Form 5021, PR-2 progress reports, PS/MMI determinations, MTUS / ACOEM alignment. The OM surface is regulatory-shaped.
Read deeper → - 08Labs
Lab interpretation patterns
Pattern recognition across canonical lab presentations (sepsis screen, AKI, thyroid panel, lipid panel) with reference-range adjustment for demographic context.
Read deeper →
Three small ways to feel the reasoning before you read the deep page.
Each capability page below ships an interactive walkthrough. They are stylised, disclosed-as-demonstration-only, and not clinical decision support — but they show the shape of the reasoning the platform produces in production.
- Interactive 01
Walk the abductive chain
Stylised case: a 78-year-old with palpitations, dizziness, and an irregular rhythm. Step through observations, hypotheses, red flags, and the ranked differential.
Open the sampler → - Interactive 02
See the vision route on three modalities
Toggle between BI-RADS, LI-RADS, and Lung-RADS schematics. See where the framework callouts live and what the structured impression looks like before the radiologist signs.
Open the demo → - Interactive 03
Twenty-four minutes versus six
Toggle between the raw clinician-typed note and the structured SOAP draft with every clause source-grounded back to the field that produced it.
Open the toggle →
Bring this list to your CMIO and CISO.
Every capability above ships in ChironAI™ today, across both editions (CDS and OM). If your evaluation needs a deeper artefact — an architecture review, a deployment RACI, a HIPAA memo — we will produce it before the engagement.