Protegrity Brings Secure Context to Enterprise AI With a Unified Policy Engine for Data, Agents and Inference

Latest Protegrity AI Team Edition release gives enterprises centralized control over how people, agents and AI applications access and use sensitive data

MENLO PARK, Calif. - Protegrity, a global leader in data and knowledge security, today announced the latest release of Protegrity AI Team Edition, introducing a new policy engine that provides a common policy foundation across data, agents and inference, three critical layers of enterprise AI.

Built on Cedar, the open-source policy language, the engine gives security teams a single solution for applying data and AI security policies to every person, agent and query, enforced at the data, agent and inference layers. It is tightly integrated with the Protegrity PII model, so visibility turns straight into enforcement.

"AI becomes more valuable when it has access to the context of the business, but that context is also where an organization's most sensitive information lives," said Michael Howard, CEO, Protegrity. "Enterprises shouldn't have to choose between AI that's useful and AI that's secure. We're building the secure context layer so they don't have to: every agent gets exactly what it needs, and does nothing it shouldn't. That's what moves AI out of experimentation and into production."

Reducing the AI Friction Tax Holding Back Enterprise AI

Enterprise AI depends on data, documents and organizational knowledge to answer questions, make decisions and act. Making that context available can also introduce security and compliance risk.

EMA's State of AI Friction research found that 83% of organizations report security or compliance reviews delaying AI into production, while 82% ship AI in a diminished state because they cannot clear the risk. Another 69% already run agents that write to databases and trigger external APIs, yet still rely on controls built for humans querying data, not agents acting on it.

Protegrity's secure context architecture addresses this challenge across data, agents and inference by protecting sensitive information, controlling access and actions based on policy, and applying guardrails to AI interactions.

New Capabilities for Securing Enterprise AI

The latest Protegrity AI Team Edition release includes:

• New policy engine: Built on Cedar and tightly integrated with the Protegrity PII model, the engine is extensible to agents and inference, not just data.

• Protegrity PII model: Protegrity's fine-tuned PII detection model achieved 0.907 precision and 0.915 recall, outperforming leading cloud, open-source and LLM-based detectors. It recognizes 69 entity types, including health conditions, biometric data and cryptocurrency wallet addresses that many detectors do not support, and includes a new benchmark dataset designed to set a higher standard for PII detection.

• Data & AI Fast Lanes: A new UI and user experience helps security teams detect and classify sensitive information, establish protection policies and apply controls before sensitive data reaches an AI model.

• AI Gateway: A centralized, extensible orchestration point enables organizations to run Protegrity guardrails alongside existing guardrails across AI workflows.

Protegrity is also making two capabilities available in tech preview:

• Secure Query: Enables transparent, high-performance search, sort, filtering and aggregate operations on protected data, with attribute-based access control and row-level filtering, without requiring triggers or UDFs.

• Iceberg Protector: Provides a secure foundation for Apache Iceberg and the compute engines and applications that access it, making the data fabric ready for enterprise AI and analytics workloads.

Building for the Agentic AI Era

This release marks the first phase of a multi-phase roadmap that will continue to expand control across enterprise AI context, including the context layer, agent intent and agent-to-agent communication. The goal is AI-ready by design, with all the context AI needs and none of the risk.

Availability

The new policy engine, Protegrity PII model, Data & AI Fast Lanes and AI Gateway are available in the latest release of Protegrity AI Team Edition. These capabilities run on the Protegrity Provisioned Cluster (PPC), the Kubernetes-based foundation that AI Team Edition deploys onto. PPC runs on Azure and AWS, and because the capabilities share one cluster, customers deploy and operate one environment in the cloud they already use. Protegrity's roadmap includes extending support to Google Cloud and to on-premises environments such as Red Hat OpenShift. Secure Query and Iceberg Protector are fully functional and already in use with customers, who are shaping the final versions before general availability.

About Protegrity

Protegrity helps enterprises protect sensitive data while putting more of it to work and as organizations move AI from experimentation into production, they need greater control over how sensitive information is accessed and used. Protegrity secures data, agents and inference for enterprise AI, enforcing one policy across all three layers so sensitive data stays usable, agents stay bounded and restricted knowledge stays protected. Enterprises use Protegrity to move more AI into production, on more of their data, with security built into each step AI takes.

Media Contact

Touchdown PR on behalf of Protegrity

[email protected]