Premera Blue Cross makes AI mainstream with Copilot and Copilot Studio

Health plans, at their best, offer a promise—to be there when life takes an unexpected turn. For Premera Blue Cross, that promise shapes everything they do, as they serve customers across Washington and Alaska with a purpose centered on making healthcare work better and values that put people first.

At Premera, innovation isn’t about novelty. It’s about enabling people to do their best work: for members, and for trust. As Dr. Adrian Mayers, Vice President, IT and Chief Information Security Officer (CISO), puts it: “Technology is here to help people.”

Getting better without slowing down

But delivering on a people-first purpose at scale takes more than intent. It takes systems and workflows that empower teams rather than slow them down, especially at a health plan, where every delay can ripple into the member experience. For Premera, that challenge was all too familiar. Manual, time-consuming work persisted across the organization, even as IT and engineering teams were already stretched to capacity.

The constraint was clear: how do you create space to experiment, learn, and modernize without asking already maxed teams to do more, while continuing to meet the demands of a highly regulated, service critical industry? At the same time, urgency was growing. As Dr. Mayers explains, what began as a desire to improve employee productivity quickly expanded into a broader set of questions: “How does this impact our customers? Is this going to enrich our value proposition? And how do we think about the competitive landscape?”

Capacity constraints were real and they directly shaped how Premera moved forward. As Carolyn Schauble, Senior Manager of Product Management, explains, the organization sought to begin learning, but engineering teams were already committed to delivering large, complex initiatives that underpin member experiences.

The answer was a different operating model. In an environment where change must be both fast and responsible, Premera leaned into its longstanding intern program, forming a cross-functional pod spanning product, platform, cybersecurity, and software development—supported by engineering managers, yet empowered to move as an agile team.

In nine weeks, the team built a range of agents: from simple tools like an acronym agent to agents that helped internal teams move faster and communicate more clearly.

Equally important was what they learned along the way. The process clarified what it would take to scale safely and showed where security foundations supported innovation, and where policies or technical pipelines needed to evolve. Today, Premera has built 900 agents. And counting.

Start where the work already happens

Microsoft met Premera where they were. “For me, honestly, it was a no-brainer. Microsoft was already a partner with us and for us. We already laid the foundation when it came to Microsoft 365 as a tenant. When it came to Azure, we were already operating within those tools and in those environments,” says Dr. Mayers.

As Schauble puts it “Premera Blue Cross is a Microsoft shop. Microsoft Word. Microsoft Teams. Microsoft SharePoint, and starting with Microsoft 365 Copilot meant meeting employees where they already work.” It also ensured trust by design. The earliest agents operated within the same permissions and boundaries employees already had. 

“These agents are really constrained to the same footprint that Microsoft 365 Copilot has. Your right to your documents follows your footprint. From there, building agents became a way to make work more repeatable: “You’re taking this big universe and making it… narrow, tighter, really repeatable,” she adds. Copilot Studio gave teams a way to move from lightweight help to workflow-specific agents, so they were not waiting for scarce engineering cycles to open.

Making AI mainstream

Premera approached making AI mainstream as an operating model—not a technology rollout. The work started with people.

Dr. Mayers describes the need to slow down to go faster, creating space for intentional training and experimentation so teams could build new capabilities without burning out.

Governance and trust were built in from day one and Mayers is clear about the non‑negotiables: “Our ability to protect, defend, enable and innovate cannot be compromised. There is no shadow AI.” Microsoft 365 Copilot aligned with that requirement by design: operating within Premera’s established identity, compliance, and data boundaries, and giving teams a sanctioned way to experiment without introducing risk.

To support that principle, Premera established a cross-functional Data, AI and Ethics Committee to set guidance, guardrails, and policies—allowing experimentation to move forward without crossing red lines. Culturally, the goal wasn’t to become “the office of no,” but “the office of how”: empowering employees to move forward together while staying true to member trust.

Just as important was making creation accessible inside the tools people already use. Copilot Studio, part Microsoft Power Platform, gave teams a single set of building blocks where automation and AI go hand in hand with a human in the loop. And because early agents stayed within existing Microsoft 365 permissions, teams could build with confidence.

As Dr. Adrian Mayers explains, “This idea of low code and no code is such a huge accessibility play. Anybody in the organization had the ability to stitch together capabilities to help their workflow, to help their productivity. That was a huge unlock.” With those barriers removed, teams were able to move quickly from curiosity to practical solutions.

Agents that help people do real work

For Schauble, the shift felt personal. The experience changed what felt possible. “My background is not in engineering. But when I built my first agent and it worked… holy moly. If I can do it, anybody can do it. Now we don’t have to just wait for the next software thing. We can get in there and solve our problems.” she says.

With that foundation in place, Premera focused on agents that addressed tangible, day-to-day challenges—starting with communication, where clarity directly affects how teams align and how members are supported. One early example was the Let’s Be Clear agent, designed to help employees write and communicate more clearly. Alongside it, the team built an acronym agent to reduce friction in a complex, highly specialized environment.

The most measurable proof point came from a high-volume operational workflow: the contract exhibit agent. A process that once took 30 to 45 minutes was reduced to about three minutes. As Schauble explains, the core pain point was manual data entry across systems. The agent uses intelligent document processing to read incoming documents, presents extracted details for human verification— “doublechecking that we got everything right”—and then routes the work to the appropriate queue, keeping a human in the loop.

Time back, better work, member value

That kind of time savings reshapes what teams can do with the same capacity. The goal wasn’t speed for speed’s sake, but rather time back: freeing people from routine, manual steps so they could focus on higher value work. As Dr. Mayers notes, Premera has “a few thousand” employees and deeply personal relationships with its members. When people need access to care or bills processed, expectations are high. Delays ripple quickly through the member experience.

In the end, Premera’s AI story is less about tools and more about culture—one shaped by the responsibilities of a health plan, where trust, privacy, and reliability matter deeply. It’s a workforce empowered to learn, protected by clear guardrails, and motivated by a mission that is fundamentally human. As Dr. Mayers describes the momentum: “People are engaged. They’re asking questions, they’re curious, they’re experimenting, they’re sharing… It’s a renaissance.”