📧 Read the full case study
Enter your email to access this case study. We'll use it to send you related content — no spam, no sharing, unsubscribe anytime.
We respect your privacy. See our Privacy Policy. Your email is stored locally in a cookie for 30 days — we do not collect it on our server.
📋 Note: This is a composite case study built from realistic deployment scenarios using our testlab infrastructure. It is not a real customer. The names, numbers, and quotes below are representative of the healthcare provider segment and are not tied to any specific organization. Real customer case studies will be published with explicit written consent.
Executive Summary
| Field | Value |
|---|---|
| Customer segment | Regional healthcare provider, 1,200 employees, 3 hospitals |
| Products deployed | AegisGate Platform Professional ($499/mo) + Lens (all staff browsers) |
| Compliance framework | HIPAA (Health Insurance Portability and Accountability Act) |
| AI tools in use | ChatGPT (clinical staff), Copilot (IT team), custom AI API (research) |
| PHI leaks prevented (first 6 months) | 134 |
| Audit outcome | PASSED with 0 findings on AI interaction controls |
| Audit prep time saved | ~40 hours (automated evidence via Trust Framework) |
| Monthly cost | $499 (Platform Professional) — Lens free for all staff |
Background
“MedHealth Regional” is a 3-hospital healthcare system with 1,200 employees. After the IT team observed clinicians using ChatGPT to draft patient communications, summarize clinical notes, and research treatment options, the Chief Compliance Officer (“Dr. Park”) raised an urgent concern:
Were clinicians accidentally sending Protected Health Information (PHI) to OpenAI?
An internal audit revealed:
- 73% of clinicians used ChatGPT at least weekly
- 41% had pasted patient-adjacent text into ChatGPT (de-identified, but inconsistently)
- 12% had pasted text containing patient names, dates of birth, or medication lists
- 0% had any automated protection
Dr. Park needed a solution before the upcoming HIPAA compliance audit — or she would need to ban AI tools entirely, which clinicians strongly opposed.
The Challenge
Dr. Park’s requirements:
- PHI detection — catch patient names, DOB, SSN, medical record numbers, medication lists, and diagnosis codes before they reach any AI service
- Browser protection — clinicians use ChatGPT in the browser, not in IDEs
- Centralized policy and audit trail — she needed evidence for the HIPAA auditor
- Air-gapped option — for the research team’s custom AI API (data residency requirement)
- No patient data stored — the security tool itself must not create a new PHI repository
- Fast deployment — the audit was in 90 days
The Deployment
Phase 1: Lens for all clinical staff (Week 1)
The IT team deployed AegisGate Lens via Chrome Enterprise policy to all 800 clinical workstations:
- Browser: Chrome (managed via Google Workspace)
- Extension: AegisGate Lens (force-installed via enterprise policy)
- Detection: PII (HIPAA subset), secrets, compliance
- Mode: Warn (clinicians see the warning and choose how to proceed)
Time: 3 days (enterprise policy rollout)
Phase 2: Platform Professional (Week 2)
The IT team deployed AegisGate Platform on an internal server:
docker run -d -p 8080:8080 \
-v ./aegisgate-data:/data \
-e AEGISGATE_COMPLIANCE_FRAMEWORKS=hipaa,gdpr \
-e AEGISGATE_AUDIT_REDACT_PII=true \
ghcr.io/aegisgatesecurity/aegisgate-platform:v4.4.0
Configuration:
- Compliance frameworks: HIPAA, GDPR (GDPR included for European patients)
- Detection policy: Block PHI (patient names, DOB, SSN, MRN, ICD-10 codes), block secrets, warn on compliance risks
- Audit logging: Enabled with PII redaction (no PHI stored in logs — only detection category and severity)
- SIEM integration: Connected to existing Splunk instance for security team visibility
- Trust Framework: Enabled for cryptographic attestation of AI interactions
Time: 4 hours (Docker deployment + Splunk integration)
Phase 3: Air-gapped Platform for research team (Week 3)
The research team’s custom AI API required data residency — no external network calls. The IT team deployed a second Platform instance on an air-gapped server:
./aegisgate-platform --mode=airgap --compliance=hipaa --audit-redact-pii
Time: 1 day (air-gapped server provisioning + Platform deployment)
Results (First 6 Months)
PHI Leak Prevention
| Metric | Before AegisGate | After AegisGate |
|---|---|---|
| PHI sent to AI services | ~22/month (estimated) | 0 |
| PHI leak near-misses caught | Unknown | 134 (blocked/warned by Lens + Platform) |
| Clinician AI usage | 73% weekly | 71% weekly (no significant change — clinicians kept using AI tools) |
| Clinician satisfaction | N/A | 89% positive (clinicians felt safer, not restricted) |
Breakdown of 134 caught incidents:
| PHI Category | Count | Example |
|---|---|---|
| Patient names + DOB | 52 | “Dear Dr. Smith, regarding John Doe (DOB: 04/15/1962)…” |
| Medical record numbers (MRN) | 31 | “MRN: 784512936 — please review the following labs…” |
| SSN | 18 | Social Security numbers in insurance pre-authorization text |
| Medication lists | 14 | Prescription details with patient identifiers |
| Diagnosis codes (ICD-10) | 12 | ICD-10 codes linked to patient names |
| Insurance information | 7 | Policy numbers, group IDs with patient names |
HIPAA Audit Results
| Audit Area | Finding |
|---|---|
| AI interaction controls | ✅ 0 findings — Platform audit logs + Trust Framework provided complete evidence |
| PHI detection and prevention | ✅ 0 findings — 134 blocked incidents documented with timestamps, categories, and redacted metadata |
| Audit trail integrity | ✅ 0 findings — Trust Framework cryptographic attestations verified |
| Data residency (research) | ✅ 0 findings — air-gapped Platform instance confirmed no external network egress |
| Staff training | ✅ 0 findings — Lens warnings served as real-time training, reinforcing the AI usage policy |
Audit prep time: 4 hours (vs. estimated 40+ hours without automated evidence)
The Audit Conversation
During the HIPAA audit, the auditor asked: “How do you ensure PHI is not transmitted to external AI services?”
Dr. Park’s response:
“Every clinical workstation has AegisGate Lens installed — it checks every prompt before it’s sent to ChatGPT and blocks PHI. Our Platform instance logs every AI interaction with PII redacted, so we have a complete audit trail without creating a new PHI repository. The Trust Framework provides cryptographic proof that the controls were active during the audit period.”
The auditor reviewed:
- Platform audit logs — 134 blocked PHI incidents with timestamps and categories (no PHI in logs)
- Trust Framework attestations — cryptographic proof of detection policy enforcement
- Lens deployment verification — Chrome Enterprise policy confirming 100% workstation coverage
- Air-gapped Platform — network diagram confirming no external egress for research AI
Result: Passed with 0 findings on AI interaction controls.
Key Takeaways
- You don’t have to ban AI tools — clinicians continued using ChatGPT at nearly the same rate (71% vs 73%). The key is detecting and blocking PHI, not blocking AI entirely.
- Real-time warnings reinforce training — each Lens warning serves as a micro-training moment. Clinicians learn what constitutes PHI through the detection feedback.
- Audit logs without PHI are critical — the Platform’s PII redaction meant the audit trail itself was not a PHI repository. This was essential for HIPAA compliance.
- The Trust Framework is a game-changer for audits — cryptographic attestation that controls were active and enforced is far more compelling than policy documents alone.
- Air-gapped deployment is feasible — the research team’s data residency requirements were met with a single binary on an isolated server.
- Cost is minimal compared to a breach — $499/month for Platform Professional vs. potential HIPAA penalties of $100-$50,000 per violation, up to $1.5 million per year for repeat violations.
Try It Yourself
- 🏢 Deploy Platform — 30-day free trial, then $499/mo Professional
- 🛡️ Install Lens for Chrome — free, force-installable via enterprise policy
- 🛡️ Install Lens for Firefox — free
- 📋 HIPAA Compliance Module — see the 43 HIPAA detection patterns
- 📊 See Pricing — compare all tiers
- 🔒 Security Overview — Trust Framework, audit logging, PII redaction
- 🏥 Live Demo — try Platform with HIPAA compliance enabled
This is a composite case study. The scenario, metrics, and quotes are representative of the regional healthcare provider segment and were validated using AegisGate’s testlab infrastructure with Keycloak OIDC, PostgreSQL, and simulated clinical workloads.