Governance teams now emphasize data, AI enablement
In some organizations, data governance programs are now focusing more on enabling employees and AI systems to use data effectively than on controlling its use.
Data governance teams are putting a friendlier face on the governance process -- and giving it a more positive business case -- by focusing more on enabling effective data use and AI applications than on usage and compliance controls.
Controlling data access and use is still a key part of governance initiatives, especially now to prevent rogue behavior by AI agents. But governance leaders, consultants and vendor executives said during Dataversity's AI Governance Online conference last week that the needle has shifted more toward making high-quality data available to both human and AI users. And there's more to it than just creating nicer-sounding internal messaging.
For example, the governance team at Children's Hospital of Philadelphia (CHOP) renamed itself the Data Trust Office (DTO) in 2024. That better reflects the DTO's primary goal of providing a centralized repository of clean, trustworthy data for broader access and use by hospital employees, said Samantha Reilly, enterprise data governance portfolio lead at CHOP.
"We don't want to be a governance program," Reilly said during a presentation. "We don't want to tell you what you can and can't do. We want to enable you."
Data catalog first, governance policy later
Anjita Shetty, the DTO's director, said expanding data access was the key focus of the data quality and governance effort from the start in 2022.
The team began not by creating a formal data governance policy, but by developing a data catalog to inventory data assets and make them more accessible to users. Shetty said leaving the policy until later in the process helped the DTO gain user trust and build stronger relationships with data owners before implementing it.
We don't want to tell you what you can and can't do. We want to enable you.
Samantha ReillyEnterprise data governance portfolio lead, Children's Hospital of Philadelphia
Initially, the team wrote just a single paragraph outlining data access rules, she added. It also built a self-service security dashboard into the data catalog as an information resource and conducted internal training on the importance of keeping data secure. "We definitely saw the need to put security at the center, despite putting the [governance] policy on the back burner," Shetty said.
But Reilly said the catalog itself became "the face of the data governance program" within CHOP, particularly as the tool's use grew. Three years ago, it contained data from two source systems and was used by about 500 developers and data analysts. As of May 2026, those figures had increased to eight data sources and a broader base of 5,500 users, which now also includes data domain owners, medical researchers and business users running self-service analytics applications.
The DTO also launched a data quality improvement effort to ensure the cataloged data was good to use, Reilly said. Steps included creating a data governance council to serve as a data-remediation advisory board and adding a data quality framework that currently contains 32 quality rules to the catalog, along with a "High Quality Data" certification badge to highlight known-good assets.
Anant Somvanshi, a senior specialist in the Data & AI Defense Office at financial services firm Vanguard, said the main focus of both data and AI governance efforts there is now on supporting innovation rather than ensuring regulatory compliance. "I've changed my hat from compliance to being an enabler within the governance function," Somvanshi said during a panel discussion.
For example, he added that a fast-path review process for AI applications with low data privacy and security risks is "an enable-friendly practice" for governance teams.
Data governance as the foundation for effective AI
Steve Holyer, a sales director for Salesforce's Informatica data management software, said during the panel discussion that when AI agents don't function properly in testing and aren't deployed for production use as a result, it's typically due to a data problem somewhere along the line.
For competitive reasons, business units often don't want to slow down AI development to put a solid data governance foundation in place, Holyer noted. "But good data governance practices enable companies to go so much faster [on AI]," he said.
The data catalog at CHOP also provides that foundation for AI, according to Shetty. She said the DTO works closely with the hospital's AI governance team, using the catalog to give AI developers a view of available data assets and models, as well as the associated context AI requires to use them accurately.
"The context that is king for AI is something we've been building into the data catalog for years," Shetty said, adding that the developers know it "can be trusted to build AI on."
Craig Stedman is a senior reporter on TechTarget's Data Technologies news team, covering analytics, data management and data backup topics. He has more than 40 years of experience as an enterprise IT writer and editor.
