Why Higher Ed IT Teams Are the Key to Building a Data-Centric Institution

“Fragmented data makes it harder to support students and partners, to improve offerings and to communicate the value of higher education,” she says.

Those gaps can delay intervention at critical moments. Indicators such as grades, billing issues and engagement often sit in separate systems, making it difficult to connect early warning signs.

Disconnected data also limits visibility into relationships with employers, research partners and other external stakeholders, where inconsistent records make alignment difficult.

Over time, that fragmentation makes it harder to track outcomes, improve processes and demonstrate institutional impact across the full student lifecycle.

Unlocking the Value of Data

Deirdre Quarnstrom, vice president of education experiences for Microsoft, says becoming truly data‑centric is less about acquiring more technology and more about unlocking the value of the data institutions already have.

“Higher education is inherently data‑rich, but too often intelligence‑poor, because information lives in silos, is inconsistently defined or can’t be trusted at the point of decision,” she says.

RESEARCH: Get the latest insight from your peers on creating frictionless experiences.

She says the real goal is a trusted foundation where insights can flow safely across the institution, enabling faster, more informed decisions today while building the institutional intelligence needed to adapt to what’s next.

“That’s how higher education delivers for the moment and prepares for the future without compromising academic values, privacy or public accountability,” Quarnstrom says.

Unifying Data Across Systems

Kew-Fickus says for IT leaders designing architectures unifying institutional data without disrupting existing systems, the best approach is to develop a central data repository — a lake house or warehouse — that can be leveraged to support multiple data use cases.

“Since higher education data is not truly ‘fast-moving,’ nightly refreshes are sufficient for most use cases,” she says.

DISCOVER: Higher education data is valuable and must be managed properly.

Quarnstrom adds that a practical approach is to modernize in place.

“IT leaders can start by identifying the highest-value, cross-functional scenarios, like connected student experiences or always-on support, then establish a shared data foundation that brings together data from student information systems, learning management systems and other services,” she explains.

From there, they can use integration patterns that respect existing investments — hybrid and multicloud connectivity, application programming interfaces and data pipelines, plus a centralized data layer — so teams can build new analytics and AI experiences on top while legacy systems continue running.

Establishing Governance

Governance is what makes data sharing possible at scale, especially in an environment with sensitive student records, research data, and regulated requirements.

“Strong governance clarifies who can access which data, for what purpose, and under what controls, so departments can collaborate without creating unmanaged risk,” Quarnstrom says.

Practically, that means consistent definitions, role-based access, auditable policies, and security that’s built in — not bolted on — so the institution can enable self-service insights while maintaining compliance, privacy and accountability across the campus.

READ MORE: Data governance is a human issue.

Kew-Fickus says data governance is about having a way to capture what is known about the data and to make informed decisions about how it is maintained, managed and used.

“When implemented well, governance helps us to create shared understanding and enable safe access,” she says.

AI Changes Data Requirements

As AI moves from isolated pilots to broader adoption, the data requirements change in three ways: quality, governance and access at speed.

Quarnstrom says institutions need well-governed, high-quality data they can trust; clear policies that preserve privacy, academic integrity and model oversight; and secure, role-based access so people (and AI systems) can use the right data without exposing what they shouldn’t.

“AI also increases the importance of a unified platform approach, so institutions can reduce tool sprawl, keep control of their data and build AI-ready foundations that support student success, research and operations with confidence,” she says.

CHECKLIST: Is your data environment ready for AI?

Kew-Fickus notes one of the hottest questions in data is how to make your data AI-ready.

She explains natively, generative AI is probabilistic while data is deterministic, making them a poor match if put together unmediated.

“The emerging solution is to pair data with contextual information, including metadata, data lineages, definitions and relationships,” she explains.

This semantic layer enables AI systems to interact with the data in a consistent and auditable way. The outputs of data governance are critical to providing this contextual layer. Governance is thus even more important in the world of AI.

“AI creates new demands, but it also creates opportunities,” Kew-Fickus says. “We are just learning how to use AI to help us make our data more AI-ready.”