Skip to main content
UnitMeshUnitMesh
Chain of Custody
Audit-Ready Evidence
Verifiable Execution

Audit-Ready AI for
Every Regulated Industry.

UnitMesh organizes AI governance around discrete, auditable decision units — each with built-in accountability (RACI), evidence-gated change management, and a complete chain of custody from input to recommendation.

Book a Demo
3
Regulated Industries
50K
Safety Run Decisions
0
Adverse Events
100%
Audit Coverage
Philosophical Foundation

Decision Units, Not Dashboards.

AI produces insights. What organizations struggle with is the last mile — translating insights into governed, auditable decisions with clear ownership and managed change.

UnitMesh is organized around David Pidsley's AI-Ready Decision Unit framework — the principle that human-AI collaboration works best when structured around discrete, auditable decisions with clear ownership and managed change. Each unit carries accountability from the moment it is defined, not as an afterthought.

Built on the philosophical and architectural concept of AI-Ready Decision Units by David Pidsley, Gartner's first Decision Intelligence Platform Leader. See also: Why Decision Intelligence Matters →

Decision

What the AI recommends

Each unit is scoped to a single, discrete decision — not a dashboard, not a score, not a report. The AI recommendation is bounded, testable, and auditable in isolation.

Ownership

RACI accountability, built in

Every decision unit ships with a RACI map — who is Responsible, Accountable, Consulted, and Informed. Accountability is structural, not procedural. It cannot be disabled.

Evidence

Chain of custody on every output

Every recommendation is wrapped in an immutable evidence record: model version, input snapshot, timestamp, and reviewer response. Audit-ready by construction, not by export.

Gartner Decision Intelligence — 5 characteristics, how UnitMesh addresses each
ConnectedCross-hospital federation — units share performance data while maintaining local governance
ContextualEach unit is scoped to a single, bounded decision — not a general model or dashboard
ContinuousShadow mode runs continuously; performance metrics trigger review and improvement cycles
CompliantImmutable evidence chain + RACI accountability + defined regulatory pathway
Cost-effectiveShadow mode activation reduces deployment risk before committing to live recommendations
PROSCI change management — also built in
Every activation follows the PROSCI ADKAR model: Awareness → Desire → Knowledge → Ability → Reinforcement. Rollout is governed, not just deployed.
ADKAR · RACI · Shadow Mode

Built for Your Industry

The same evidence infrastructure. The compliance language your regulators already speak.

Regulated Industry — Healthcare
  • Clinical AI decisions cannot be audited after the fact — no evidence trail, no regulator defense.
  • Model performance degrades silently — no systematic monitoring between audits.
  • EHR integration requires point-in-time data that can be reconstructed on demand.

UnitMesh activates clinical decision units through a governed shadow-to-active pathway. Every recommendation carries a full evidence chain: model version, feature snapshot, timestamp, and clinician response — all linked by deterministic evidence identifiers. When a regulator asks "why did the AI say that?" you have a complete answer.

AI governance where every clinical decision traces back to verifiable evidence.
50K Simulated Decisions0 Adverse Events4 Hospital Scenarios
FDA De Novo — target DEN230156 (2029)CE Mark — target 2027ISO 13485 — in progress

The Four Clinical Decision Units

Each unit runs in shadow mode, passes formal audit, and requires dual or triple sign-off before going live.

🏥

CHF Discharge Routing

Evidence-based discharge tier recommendation for congestive heart failure patients — reducing preventable readmissions.

💊

Medication Reconciliation

Automated reconciliation routing at transitions of care, flagging high-risk discrepancies for pharmacist review.

📈

ICU Readmission Risk

Post-ICU readmission probability scoring, enabling proactive step-down and early intervention.

Sepsis Early Warning

Continuous sepsis risk scoring with THRIL-composite alert routing to nursing and rapid response teams.

How Activation Works

Every decision unit follows the same governed pathway before a single recommendation reaches a reviewer.

01

Shadow Mode

RACI accountability mapping

90-day observation under live conditions. The AI generates recommendations in parallel with reviewers — no operational impact. RACI ownership is established before a single live recommendation is issued.

02

Formal Audit

Evidence-gated approval

Shadow period evidence is reviewed against threshold criteria. Dual sign-off required (Red-tier units add Legal). No unit advances without a closed evidence chain — the gate is structural, not procedural.

03

Active Deployment

PROSCI ADKAR rollout

Recommendations surface via FHIR CDS Hooks. Rollout follows the PROSCI ADKAR model — Awareness through Reinforcement. Every decision is logged with full provenance: model version, features, timestamp, reviewer response.

Healthcare Regulatory Pathway

UnitMesh is simulation-validated and tracking a defined regulatory pathway. Insurance/Climate and Pharma are in evaluation-stage partnerships.

Target: 2027
CE Mark
Medical Device Regulation (EU) 2017/745
Target: 2029
FDA De Novo
DEN230156 — Product code QMV
In Progress
ISO 13485
Medical Device Quality Management System
0 recorded
Adverse Events
In 50K-decision safety validation run
Healthcare Simulation Reference

3IT Health Network

3IT Health is the simulation reference network used to validate UNIT MESH — four decision units running across four hospital scenarios. The safety validation run completed 50K decisions with 0 adverse events and demonstrated federated model improvement across activation stages.

The trust layer powering every decision — ProvenLake — is the same infrastructure available to any regulated-industry application built on 3it.dev.

4
Hospital Scenarios
16
Unit Activations
50K
Simulated Decisions
0
Adverse Events

Pricing

Start free and grow into full regulatory-grade evidence chains as your deployment matures.

Evaluation
Explore the platform
Free
  • Single industry vertical
  • Up to 500 verified decisions
  • 30-day evidence retention
  • Internal team verification only
  • 2 governed pipelines
  • Community support
  • Full chain-of-custody logging
Start Evaluation
Trust
Ongoing audit readiness
Contact Sales
  • Single industry vertical
  • Up to 10,000 verified decisions / month
  • 12-month evidence retention
  • Internal audit access
  • Up to 5 governed pipelines
  • Email support (2-business-day SLA)
  • Decision Registry access
  • Performance monitoring dashboard
Contact Sales
Most Popular
Compliance
Regulator-ready evidence chains
Contact Sales
  • Up to 3 industry verticals
  • Unlimited verified decisions
  • 36-month evidence retention
  • External auditor access portal
  • Unlimited governed pipelines
  • Priority support (next-business-day SLA)
  • Formal audit workflow support
  • Multi-approver sign-off chains
  • Custom evidence retention policy
Contact Sales
Enterprise
Multi-industry, multi-region
Contact Sales
  • All industry verticals
  • Unlimited verified decisions
  • Custom retention (up to 10 years)
  • Regulator-facing evidence export
  • Unlimited governed pipelines
  • Dedicated support with named CSM
  • On-premise or private cloud deployment
  • Custom approval chain configuration
  • Federated multi-region deployment
  • SLA negotiated per contract
Contact Sales

Book a Demo

See UnitMesh in action. We'll reach out within one business day.