Enterprise architecture needs to let go of rigid 'gatekeeping' to stay relevant in the AI era

Enterprise architecture needs to let go of rigid 'gatekeeping' to stay relevant in the AI era
EDB News Desk - May 23, 2025
ac8c7b66-ab6a-4993-887b-8698fc55bed6.jpeg
Credit: Outlever

KEY POINTS

  • Enterprise architecture is under pressure to transform from a control-focused role to a strategic enabler in response to agile and AI-driven business demands.

  • Ryan Pehrson advocates for EA to function like an air traffic controller, providing oversight without hindering fast-moving product teams.

  • The shift to a decentralized model requires a cultural change, integrating transparent decision-making and real-time data into product team workflows.

"The world has changed. Enterprise architecture really hasn't. There's friction between teams that are trying to move fast and adopt new technology running up against traditional control structures."

1247f5c1-d316-476f-ac50-5f3878667d81.jpeg
Global Technology Leader, Enterprise Cloud & Infrastructure
Ryan Pehrson

Enterprise architecture is showing its age. Built for a slower, more centralized era, its traditional frameworks are straining under the weight of agile teams, AI initiatives, and real-time business demands. To stay useful, EA must shed its static past and evolve into a living, adaptive function.

Ryan Pehrson is a Global Technology Leader with a track record of leading digital transformations in complex, high-stakes environments. With deep enterprise architecture experience—including for a top 25 biopharma firm—he sees firsthand why EA must evolve or risk fading into irrelevance.

Stuck in the past: "The world has changed. Enterprise architecture really hasn't," Pehrson says. "There's friction between teams that are trying to move fast and adopt new technology running up against traditional control structures." EA’s legacy role as a gatekeeper—intended to preserve strategic coherence—now too often "slows everything down." And when that happens, Pehrson warns, "EA becomes irrelevant. People will just go around it."

Ready for takeoff: To stay relevant, EA needs to shift from gatekeeper to enabler. As Pehrson puts it, EA should be "an air traffic controller with an eye in the sky view and strategic oversight" without grabbing the controls. That means building a living map and boosting observability, ensuring product teams stay visible and aligned without being slowed down. Ultimately, "EA has to evolve to be more like a conductor of a very decentralized, fast moving model," explains Pehrson.

Culture is key: Traditional EA tools are falling short. Designed for diagrams and metamodels, they don’t fit today’s fast-moving, code-first product teams. "How can you bake it into everything that a product team does?" Pehrson asks. The answer lies in transparent decision records and living data from instrumented systems—but none of it sticks without culture. "If you don't implement a culture where product teams have to keep that information up to date, then it fails," he explains.

"EA has to evolve to be more like a conductor of a very decentralized, fast moving model."

1247f5c1-d316-476f-ac50-5f3878667d81.jpeg
Global Technology Leader, Enterprise Cloud & Infrastructure
Ryan Pehrson

Value, not vanity: Beyond tools and culture, Pehrson reframes EA’s role not as tech oversight, but as a true driver of business value. "Where EA really shines is to deeply understand how the business operates, the value streams, how we produce value for the customer," he explains. "Then understanding how all the technology components relate to that." The goal: "What levers can we pull in order to get more value for the customer and therefore create more enterprise value?" says Pehrson.

Dream or nightmare?: "The evolution of AI in the enterprise is exciting," Pehrson says. "But it’s also what keeps me up at night." The risk isn’t the tech itself, but the speed and sloppiness with which it’s often deployed. "Those who step back, take a strategic approach, and then move are going to be more successful," he advises. That means getting serious about data quality and governance. Without both, AI becomes a liability, not a leap forward.

It takes a village: "We have to be better than we have been in order to make AI really fly," Pehrson says. That means breaking down silos and building tighter alignment: "Enterprise architecture, product teams, and platform teams need to work together in a new way in order to truly get enterprise value and limit the risks," says Pehrson.