• CLAIRE Agents help cut down manual data work and reduce IT backlogs.
• They automate data tasks so teams spend less time fixing and more time using data.

Managing enterprise data is a messy job. Most companies have too much of it, spread across too many systems, and rely heavily on IT teams to make sense of it all. Informatica’s CLAIRE Agents are designed to ease that burden by introducing goal-driven, autonomous agents that can handle time-consuming data tasks more efficiently.

In an interview, Sumeet Agarwal and Gaurav Pathak—both Vice Presidents of Product Management at Informatica—explained how CLAIRE Agents are built to reduce manual data work, automate common tasks, and increase trust in AI-driven processes.

What makes Informatica’s approach different, according to Agarwal, is a focus on building accurate agents that are grounded in trusted enterprise data, can work across business systems, and are managed with clear controls. “Unlike many fragmented or rule-based AI agent frameworks,” he said, “Informatica emphasises grounding AI agents in trusted, high-quality enterprise data to enhance decision-making and reduce errors.”

Many agent platforms, he said, still struggle with scaling and security because they rely on patchwork integrations. Informatica’s model is built with enterprise controls in mind from the start.

Gaurav Pathak, VP of Product Management for AI and metadata, believes CLAIRE Agents are well suited to address pain points that current AI tools don’t handle well. “Many organisations perform data management tasks manually using code, without a central metadata knowledge base,” he explained. That creates slow, fragile systems that are hard to maintain.

One of the biggest problems, he said, is the backlog of IT requests that comes from trying to build and maintain all this code. “This results in departmental and shadow-IT efforts, leading to more fragmentation, rework, and non-standardised output.”

CLAIRE Agents are designed to reduce that backlog. Users can type a natural language prompt to search for existing data products, or describe a new data need. The agents then find the right sources, prepare the data, and document the process. “Productivity improvements from CLAIRE Agents are expected to be exponential,” Pathak said, “as data product developers delegate more and more goals.”

They can also clean up existing data by checking for errors, tracing data lineage, or fixing broken pipelines. Because the agents work across platforms like AWS, Azure, Google Cloud, Oracle, and Salesforce, they can pull context from across the business.

Security and trust are still central concerns. CLAIRE Agents only operate within sandboxes controlled by Informatica’s platform. “CLAIRE Agents always require human-in-loop for running any data management code or jobs by default,” said Pathak. That default can be adjusted over time, but the system keeps a clear permission structure. Organisations can also add their own rules or limits to guide agent behaviour.

Agarwal said the governance model includes role-based access control, session management, and policies around authentication and rate limits. “This emphasis on responsible AI, powered by high-quality data, is critical for trust and adherence to regulations.”

For businesses that want to expand AI use beyond technical teams, no-code orchestration is a key part of the plan. “No-code or low-code agent orchestration is pivotal in democratising AI adoption,” Agarwal said. It allows people in business units—not just developers—to set up and run their own agent workflows. The goal is to increase access and reduce reliance on IT for every task involving data.

This cross-functional approach also helps different teams share results and avoid doing the same work twice. Instead of data being siloed or recreated from scratch, it becomes more reusable and discoverable across departments.

To track how CLAIRE Agents are performing, Informatica looks at both usage and output. Pathak said adoption is measured through task completion rates, reduction in manual work, and how often agents are used. Reliability is monitored through error rates and user feedback. “Feedback from these metrics informs the Agentic AI roadmap by highlighting areas for improvement in agent autonomy, adaptability, user experience, and integration capabilities,” he said.

As CLAIRE Agents are expected to run across a mix of cloud and on-prem systems, Informatica has tried to make sure they behave consistently regardless of the environment. The agents are built on a common metadata layer that gives them the context they need no matter where data lives. They also support open standards, including the Model Context Protocol (MCP), to help them plug into other platforms more easily.

In addition to technical design, Agarwal pointed to the company’s work with cloud providers like AWS, Azure, and Salesforce. These integrations, he said, help ensure the agents can operate efficiently across different systems without losing performance or trust.

The goal is not to replace people, but to help them do more. As Pathak put it, CLAIRE Agents are there to “significantly boost productivity by automating complex data preparation and management tasks.” That includes finding what data exists, building new products, and improving what’s already in place—so teams can spend less time chasing problems and more time using data to make decisions.