{"id":"mcp-data-residency-insurance","slug":"mcp-data-residency-insurance-compliance","title":"Data Residency, HIPAA, and MCP: What Compliance Officers Need to Know Before Connecting AI Agents to a CRM","excerpt":"Letting Claude read PHI sounds scary until you understand the architecture. Here's the compliance officer's guide to MCP — and the contractual structure that keeps it audit-clean.","content":"<p>Insurance CRMs hold PHI. Connecting AI agents to PHI without thinking through the compliance architecture is reckless. Done correctly, MCP is no riskier than any other API integration.</p>\n\n<h2>The Three Questions Compliance Always Asks</h2>\n<h3>1. Where Is the Data?</h3>\n<p>MCP doesn't move data anywhere new. The agent calls the CRM API. The response goes to the LLM provider. From there, it depends on the LLM contract.</p>\n\n<h3>2. Does the LLM Provider Sign a BAA?</h3>\n<p>Anthropic offers a BAA for Claude Enterprise. OpenAI offers BAAs for ChatGPT Enterprise and API customers. If you require HIPAA, sign a BAA with whichever provider your team uses.</p>\n\n<h3>3. Are Logs Auditable?</h3>\n<p>unLocked CRM logs every API and MCP call with X-Request-Id, timestamp, key fingerprint, and resource. Combined with your LLM provider's logs, you have full call traceability for any audit.</p>\n\n<h2>Practical Controls</h2>\n<ul>\n<li>Use scoped API keys per agent type</li>\n<li>Disable PHI fields for agents that don't need them</li>\n<li>Mask SSN, DOB, and Member ID in MCP responses for analyst agents</li>\n<li>Require approval for any write operation (Claude can draft; CSR sends)</li>\n<li>Monthly access review of all issued keys</li>\n</ul>\n\n<h2>What unLocked Does Out of the Box</h2>\n<p>Field-level masking, scoped keys, audit logs, BAA-eligible architecture. Compliance review of MCP integration takes hours, not weeks.</p>","category":"Compliance","featureLink":"/mcp-server","cluster":"mcp-api","date":"2026-11-18","lastModified":"2026-11-18","readTime":"6 min","author":"unLocked Team"}