ON-DEMAND SHOWCASE
PII Detection, Service Accounts, RBAC, and What's Next for Production AI
Enterprise AI projects don't stall at the model layer; they stall when governance breaks down at scale. When sensitive data crosses boundaries unchecked, when agents mask human credentials, and when no one can answer a compliance team's questions about what ran, when, and why.
Watch an on-demand walkthrough of Connect AI’s new, HIPAA-compliant feature set: every new capability, with real-world workflows, product demos, and answers to questions from the team.
What you’ll learn
The showcase covers five new capabilities, demoed in the product:
HIPAA-compliant architecture – Connect AI can now connect EHRs, payer systems, analytics platforms, and enterprise applications with the governed architecture, documented ePHI flows, and audit trails that compliance requires.
PII detection and token masking — Enforce configurable warn, redact, or block policies at the MCP tool-call boundary, inbound and outbound, before sensitive data reaches an AI model or returns to a user.
Service accounts for agents — Give every autonomous process its own scoped, auditable machine identity. Agents and scheduled pipelines connect without requiring a human OAuth session, and every action they take is independently traceable.
Role-Based Access Control (RBAC) — Four universal platform roles plus custom role definitions with additive permissions, scoped to specific connections and workspaces. A second governance layer on top of source-inherited permissions.
Connect AI skills — Downloadable plugins, published on GitHub, that guide Claude, ChatGPT, Gemini, and other LLMs through the correct source discovery, reducing token overhead and eliminating first-call failures.
Speakers
Padma Chilakapati, Director of Product Management, Security & Governance — CData
Matt Dallhoff, Principal Technical Product Marketing Manager — CData
Cameron Leblanc, Senior Technology Evangelist — CData
