Protegrity has released an updated version of Protegrity AI Team Edition with a unified policy engine that controls what people, AI agents and applications can access and do.
Built on the open-source Cedar policy language, the engine applies data and AI security policies to users, agents and queries. It integrates with the Protegrity personally identifiable information model, allowing security teams to apply controls based on the sensitive information the model detects.
“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.”
Enterprise AI applications use internal data, documents and company knowledge to answer questions and take action. Providing access to that information can create security and compliance risks, particularly when agents can write to databases or trigger external application programming interfaces.
EMA’s State of AI Friction research found that security or compliance reviews have delayed AI production deployments at 83% of organizations. The research also found that 82% have deployed AI with reduced functionality because they could not resolve the associated risks.
Another 69% of organizations already operate agents that can write to databases and trigger external APIs, according to the research. Many continue to rely on controls designed for people retrieving data rather than agents using that data to complete tasks.
Protegrity’s architecture applies controls across the sensitive data supplied to AI systems, the permissions assigned to people and agents, and each prompt and response. The new policy engine determines what each user, agent and query can access and which actions they can perform.
The release also includes Protegrity’s fine-tuned PII detection model, which recognizes 69 types of sensitive information. Supported categories include health conditions, biometric data and cryptocurrency wallet addresses.
Protegrity reported that the model achieved precision of 0.902 and recall of 0.910. The company said those scores were higher than results from OpenAI’s Privacy Filter, Amazon Comprehend, Microsoft Presidio and GLiNER.
A new Data & AI Fast Lanes interface helps security teams detect and classify sensitive information, create protection policies and apply controls before the information reaches an AI model.
Protegrity has also introduced AI Gateway, a centralized orchestration point that checks prompts and responses throughout AI workflows. Organizations can use it to run Protegrity’s guardrails alongside their existing controls.
Secure Query and Iceberg Protector are available in technology preview. Secure Query supports search, sorting, filtering and aggregate operations on protected data. It also provides attribute-based access control and row-level filtering without requiring database triggers or user-defined functions.
Iceberg Protector extends data protection to Apache Iceberg and the compute engines and applications that access it. Protegrity said Secure Query and Iceberg Protector are fully functional and already in use by customers, whose feedback will help shape the final versions before general availability.
The release is the first phase of a broader roadmap that will extend Protegrity’s secure context controls to the information supplied to AI systems, agent intent and communication between agents.
The policy engine, PII model, Data & AI Fast Lanes and AI Gateway are available in the latest version of Protegrity AI Team Edition. The capabilities run on Protegrity Provisioned Cluster, a Kubernetes-based foundation available on Microsoft Azure and Amazon Web Services.
The shared cluster allows customers to deploy and operate the capabilities in one cloud environment. Protegrity plans to add support for Google Cloud and on-premises environments such as Red Hat OpenShift.
Join our LinkedIn group Information Security Community!
