Informatica Inc. is turning up the volume on its agentic artificial intelligence push with today’s announcement of a series of updates to its Intelligent Data Management Cloud, as the company positions AI agents as the next step in helping enterprises manage and act on data.

The Fall release introduces a suite of purpose-built AI agents under Informatica’s Contextual Learning AI and Reasoning metadata intelligence engine, along with a set of engineering tools designed to help customers create and orchestrate their own autonomous data agents. The idea is to make data management something that can be reasoned with, automated and made smarter without extensive coding.

The Fall update brings the first wave of Claire Agents into general availability, including a Data Exploration Agent for master data management, an Enterprise Discovery Agent, an Extract/Load/Transform Agent and a Product Help Agent. A Data Quality Agent is now in public preview, and a Product Experience Agent is in private preview.

Each is accessible via Informatica’s Copilot, which uses a natural language interface. The goal is to reduce the time data engineers spend on routine tasks and entrust more sophisticated data operations to business users, said Chief Product Officer Krish Vitaldevara.

“These agents bring together the different kinds of data management tasks, make them available easily through prompts, so you can either directly interact with them, work through copilots or invoke them through the other agents,” he said. “Every capability of IDMC is going to be available through agents in the fullness of time.”

Enhanced natural language planning and reasoning capabilities allow agents to go beyond reporting a problem to suggesting ways to solve it. “If you ask about the quality of a data set, the system can now reason through the context and propose fixes,” Vitaldevara said. “It’s an evolution from rules-based workflows to reasoning-based outcomes.”

The agents are capable of taking action, but “there’s always a human in the loop,” Vitaldevara said. “Agents can recommend corrections, and you decide whether to apply them and automate them for future use.”

Build your own agents

Informatica is also launching tools for enterprises to roll their own agents. AI Agent Engineering is a no-code development canvas for building, orchestrating and governing AI agents tailored to unique workflows. It lets developers bring their own models, vector databases and domain logic to craft custom agents and Model Context Protocol clients and servers. Informatica said the toolkit can reduce agent development time from weeks to minutes.

Also new is AI Agent Hub, a centralized directory where pre-built agents and automation recipes can be shared, validated and reused across teams.

A new security feature in the Fall release is controlled unmasking. It allows authorized users to view masked or protected data under policy-driven conditions without compromising overall privacy or regulatory compliance.

“We are investing in both the understanding of unstructured data, what metadata to extract and what new masking rules should look like for the unstructured data,” Vitaldevara said. Multifactor authentication is also now available for all native IDMC accounts.

Vitaldevara said Informatica’s preview programs for the new tools have been “significantly oversubscribed. We’re seeing customers with triple-digit agent counts already,” he said. “Just like with tasks and data assets, the number of agents is going to explode.”

Image: SiliconANGLE/Ideogram