By Katie Anderson, Chief Editor, Pharmaceutical Online

Artificial intelligence has no shortage of promise in pharmaceutical operations. The harder question is how to move from promising ideas to practical, compliant, and adopted solutions. For Michael Grischeau, Director of Data Analytics and Management Review at AbbVie Inc., and Paula Gamboa, Associate Director of Data Strategy and Innovation at AbbVie, the answer begins with cross-functional collaboration.

In their discussion of AbbVie’s Operations Quality Assurance Innovation Accelerator at the ISPE AI in Life Science Summit, Grischeau and Gamboa described a framework designed to help operations quality teams identify the right problems, assess readiness, and advance innovation in a way that supports both business value and regulatory expectations. The initiative brings together governance, leadership, data strategy, and community-based problem solving to accelerate ideas without losing sight of compliance.

From AI Excitement To AI Readiness

The team used a simple analogy to explain the challenge: building with a Lego set. Sometimes you just want to get it done so that you can enjoy the final result, but the framework, the components, and the process matter. Similarly, a team may have a clear picture of the final result, but success depends on having the right pieces, instructions, and process. “We realized we needed to develop a process around the whole concept of AI,” explained Grischeau. Organizations may have use cases and a compelling vision, but without a framework for intake, prioritization, governance, data quality, and adoption, the pieces do not fit together.

“We realized AI wasn’t really the hard thing; everything around it was,” Gamboa added. Teams were quick to propose AI solutions, but the more important question was whether the organization was ready to implement them responsibly and sustainably.

That readiness gap is familiar across regulated environments. Poor data quality, resistance to change, low adoption, disconnected workflows, and unclear AI governance can all prevent a promising tool from becoming a useful capability. As Grischeau explained, even well-built technology can fail if users return to familiar spreadsheets because the solution does not fit their daily work.

Gamboa clarified that the barriers to integrating AI solutions are often not technological. “They are readiness problems, and they are people problems,” she noted. That realization pushed the team to rethink how innovation should be built inside operations quality.

Building A Community Before Building Solutions

AbbVie’s approach began by creating a community within operations quality. Today, that community includes more than 120 cross-functional members across 11 functions. The sessions are designed not only to share updates, but also to educate participants, surface problems, and connect people who may be solving similar challenges in different areas of the business.

That community of practice (COP) became the people infrastructure for transformation. When one member raised a challenge, another could share how they approached a similar problem. Instead of thinking in silos, teams began to recognize patterns, reuse capabilities, and build on one another’s work. “We are thinking in a much more cross-functional way. Just sharing information wasn’t enough. We wanted to start bringing things to action and really start thinking about how we build on what everybody else is doing,” added Gamboa.

Grischeau continued that with the people component in place, they really just needed the framework to bring innovation to reality, and they is where the Innovation Accelerator came into play.

The Innovation Accelerator Framework

The Innovation Accelerator gave teams a structured way to bring ideas forward, evaluate opportunities, and develop solutions with greater discipline. It provided tools, templates, and guiding questions that helped teams define the problem before jumping to the technology.

Instead of asking only, “What AI tool do we want?” teams were encouraged to ask: How much time is being spent on the current process? What pain point are we solving? What is the business value? What data do we have? What governance is required? Who will own the work, and who will adopt it?

In 2025, the team pitched the framework to leadership with a goal of flipping the traditional way of working. “Leadership would often say, these are the things I want you to work on. We wanted to flip that to crowdsource these ideas from the people who are living from the pain and the problem,” continued Grischeau.

The response was significant: the team received approximately 100 opportunities, which they then narrowed the list to 10 priorities. To their surprise, 58 volunteers from multiple functions volunteered to help move the work forward. The principle guiding the work was “progress over perfection.”

What Happens When Collaboration Gets Hard

The Accelerator also revealed a familiar truth that cross-functional work can feel slow. Teams encountered challenges such as unclear ownership, limited sponsorship, difficulty sustaining momentum, and the complexity of assigning accountability for problems that cut across functions.

Rather than viewing those hurdles as failures, the team used them as signals. They stepped back to understand what each team needed, where projects overlapped, and where existing capabilities could be combined or adapted. This helped reduce duplication and encouraged teams to stop reinventing the wheel.

Just as importantly, participants were developing skills that extend beyond any individual project: critical thinking, project management, communication, storytelling, and conflict resolution. For Grischeau, there was so much value in those intangibles. “It is less about data science and life science and more about people science. That has been our experience,” he concluded.