What happened
On Nov 3, 2026 Protegrity announced the latest AI Team Edition release, adding a unified policy engine built on Cedar that applies one set of policies to people, agents, and every query; a fine-tuned PII model (0.902 precision / 0.910 recall, 69 entity types) integrated so visibility turns into enforcement; Data & AI Fast Lanes; and an AI Gateway that runs Protegrity guardrails alongside existing ones on every prompt/response. Secure Query and Iceberg Protector are in tech preview.
Why it matters
This extends an established data-security platform from protecting data to governing the agentic layer — a single policy engine deciding what agents and inference calls can see and do, which directly reduces the 'AI friction tax' (83% of orgs report security/compliance reviews delaying AI) and the risk of agents acting on sensitive data with human-era controls.
Applicability
Data-security and platform teams at regulated enterprises (finance, health) using AWS/Azure should evaluate the unified policy engine and AI Gateway for prompt/response control and PII detection before letting agents act on protected data.