What data should autonomous AI agents be allowed to access? Immuta has expanded the functionality of its flagship platform to include data governance and provisioning capabilities for agentic AI.
AI agents need data to do their work, just as human workers do. But providing those agents with access to the right data with the right safeguards can be a challenge.
Immuta has added new capabilities to its data platform that the company says enable businesses and organizations to provision and govern AI agent access to enterprise data in real time.
“Agents break the old data governance models,” said Immuta CEO Matt Carroll (pictured) in an interview with CRN.
[Related: 12 Agentic AI Startups To Watch In 2026]
Decisions about what data can be accessed by which employees have traditionally been made by chief data officers, data stewards, data custodians, and other such managers—a process that could take hours or days.
“But agents are different. They work in seconds and minutes and operate across all your data,” Carroll said. “Hours and days are unacceptable.”
Immuta, founded in 2015 and based in Boston, develops a data governance and provisioning platform, with a data policy engine at its core, designed to streamline the process of providing business workers with access to the data they need for their jobs—but only the data they need.
The platform is used to establish purpose-based data governance policies and then orchestrate and automate all aspects of data provisioning, including monitoring data usage and enforcing access privileges, to reduce the risk of data misuse.
In Immuta’s early days the platform provided data access capabilities for business analysts and other business users. Those needs accelerated with the wave of generative AI adoption. “When generative AI took hold in the market, everyone became a data consumer,” Carroll said, noting that organizations “needed to deliver governed data at scale.”
Immuta has been rapidly adapting its platform to meet the needs of the AI world. In June 2025 it added new AI capabilities to the platform’s data provisioning functionality, helping organizations more quickly provide the right data to the right users “with appropriate policy enforcement and full auditability,” according to the announcement.
And Immuta further expanded the data provisioning platform in October with new Guardrail Policies, Policy Exception Workflows and Multi-Approver Workflows capabilities.
Carroll said the development and implementation of AI agents, starting in mid-2025, happened surprisingly quickly.
“Agents break the old data governance models,” he said. Traditional manual approvals and ticket-driven workflows are inadequate for agents, which need access to the same data that humans do but need it almost instantly without waiting for approval from data stewards.
New Agent Data Access Functionality
The addition of the Agentic Data Access module to the Immuta platform addresses the issue by treating agents as what Carroll described as “first-class data users,” like humans, while at the same time limiting their access to just the data they need for specific tasks, known as “least access privileges.”
Using the new Agentic Data Access capabilities, data managers can establish data access policies and entitlements for AI agents that are either acting independently or on behalf of a human worker.
Immuta modified the data govern, access and comply capabilities within the Immuta platform’s policy engine, Carroll said, providing a data “authorization and delivery layer” for agents. The “zero standing privileges” functionality automatically removes data access authorization once a task is complete.
Agentic Data Access also includes a registration system for agents and audit trail capabilities that record agents’ data access activities.
Later this year Immuta will expand Agentic Data Access with “semantic governance” capabilities that alert agents to data that might help them in their tasks and immediately establish governance policies for that data. The platform will also allow agents to independently initiate data access requests.
Channel Opportunities
In the channel, Immuta partners with consulting firms and with regional and global systems integrators who work with businesses and organizations to implement the Immuta platform and develop data governance practices.
Carroll said the new Agentic Data Access creates an opportunity for partners to provide customers with new consulting services. “Every company needs guidance on data access policies for agents,” he said.
More broadly, he said businesses and organizations need to retrofit their legacy data infrastructures to accommodate rapidly proliferating AI agents.
“I think there is a huge opportunity for global SIs and regional SIs to deliver data for agentic systems and make sure [data processes] are optimized for speed and scale for agentic architectures,” the CEO said.
