Walk one note through your workflow: clinician records, AI drafts, CDS reviews, EHR entry, sign-off. GLACIS turns each handoff into cryptographic evidence.
The Scenario
The 98% edit-free claim is compelling. Procurement wants evidence:
Compliance: How do you prove CDS actually reviewed each note—not just clicked approve?
Legal: If a note is disputed, can you reconstruct what the AI drafted vs. what CDS changed?
CMO: How do we verify the signed version matches what went into the EHR?
Your audit logs say "reviewed."
Procurement says "prove the human actually looked."
The Questions
Patient challenges note accuracy. Was this AI hallucination or CDS error? Without receipts, you can't prove who authored what—or when.
Competitors add CDS-like features and make the same marketing claims. Your 15 years of experience becomes hard to prove.
The opportunity: You actually do have human oversight. GLACIS lets you prove it cryptographically on every note.
The Walkthrough
We attach at each step of your workflow: audio capture, AI draft, CDS review, EHR entry. Each handoff generates a cryptographic receipt—proving what happened, when, and by whom.
Every note has a receipt chain. You can show exactly what the AI drafted, what the CDS changed, and when the clinician signed. No more "trust us."
Competitive Moat
Anyone can claim CDS review. Only one company can prove it cryptographically—on every note.
COMPETITIVE ADVANTAGE
Prove human review: Every note has attested CDS touchpoints
Win disputes: Reconstruct what happened, when, by whom
Defend your moat: Competitors can't copy cryptographic proof
REGULATORY DEADLINES
Standard of care precedes law. NIST AI RMF and ISO 42001 are already the de facto duty of care.
The window: In 18 months, AI vendors without cryptographic proof will lose deals to those who have it.
The Insight
Patient says the note is wrong. The question isn't "did someone review it?"—it's "can you prove it?"
Without GLACIS
Dispute investigation
"Our logs say CDS reviewed it"
Problem: You wrote the logs. They're contestable.
Procurement Asks
"Prove the human actually reviewed"
Show timestamps, diffs, signatures
Result: Deal stalls. Same questions every time.
With GLACIS
Show the receipt chain
AI draft → CDS review → EHR → sign-off
Third-party witnessed. Immutable.
Live Demo
Every AI request gets attested. Third-party witnessed. Zero data egress. Sub-50ms.
Proof Artifacts
Live at app.glacis.io — verify receipts, inspect witness proofs, and map evidence to governance frameworks.
One encounter. Proof chain. Verifiable without PHI exposure.
NIST AI RMF • ISO 42001 • EU AI Act • Colorado
70+
Patent claims filed
4 families • Fenwick & West
Defensibility
70+
Patent claims filed
4 families • Fenwick & West
Co-Epoch Attestation
Binds receipts to specific binary versions. Prevents "binary substitution" attacks.
Statistical Safety Signals
Mathematically rigorous sampling for continuous compliance verification.
Zero-Payload Egress
Cryptographic commitments without data exposure. PHI never leaves boundary.
Federated Witness Network
Distributed attestation for independent verification at scale.
Filed Nov 2025 — First-mover advantage in attestation-based AI compliance.
Next Steps
Your hybrid model is exactly the differentiator that cryptographic proof can make defensible. Let's see where it fits.
WHAT A PILOT MIGHT LOOK LIKE
→ Instrument your workflow at key handoffs
→ Generate receipts for CDS review events
→ Build a sample audit package for procurement
→ Evaluate fit before full integration
The Evidence Layer for AI
Appendix
ProScribe Pipeline
Audio → AI → CDS → EHR
GLACIS
SIDECARAttest → Hash → Sign
<50ms • 0 bytes PHI out
Witness Network
Hashes only
Merkle anchor
T+0s Audio captured
Clinician records encounter. Audio is hashed locally and bound to the session ID.
T+180ms AI draft generated
Your model generates the note. Input and output are hashed before handoff to CDS.
T+4m CDS review attested
CDS opens draft, reviews, and submits. Session duration and edit delta are captured.
T+5m EHR entry verified
Final note is hashed at EHR submission. Chain links draft → review → submission.
T+8h Clinician sign-off
Clinician reviews and signs. Receipt chain is complete and anchored in witness network.
Zero-Egress = No BAA
No PHI leaves customer boundary. Eliminates months of legal review.
Third-Party Verifiable
Ed25519 signatures + RFC 6962 Merkle trees. Independent verification.
Drop-In Deployment
Lambda layer or container sidecar. Works with any LLM provider.