Somewhere in nearly every major pharma and medtech organization, there is a slide deck showing dozens of active AI initiatives—and a CFO asking why so few are showing up in the financials.

That question is becoming harder to avoid. Our survey of 133 life sciences executives, run in partnership with the Mayfield Fund, found that only 20% of organizations are consistently deploying AI at scale and achieving meaningful, measurable value. The tools have not failed the other 80%. Their structures, behaviors, and ambitions have not kept pace. And the distance between them is about to widen.

Figure 1

Agentic AI—systems capable of autonomous reasoning and multi-step execution—is moving quickly into enterprise use. Companies that have built the organizational muscle to scale AI are now first to capture its next wave. Those still in pilot mode will find it increasingly hard to catch up.

Successful scalers get three things right:

• a transformative ambition grounded in a small number of strategic bets;
• the resolve to reimagine entire workflows from outcomes rather than optimizing individual steps; and
• an organizational operating model that accelerates adoption by embedding the behavior change and talent required for AI to take root.

Transformative visions grounded in strategic bets

Most life sciences organizations do not lack AI activity. But many lack focus. Successful scalers anchor on a small number of transformative bets tied to business outcomes, rather than spreading resources across disconnected use cases.

Take one major medtech firm that committed to over 100 AI initiatives, yet projected just around 2% EBIT improvement. The real shift came when leadership stopped asking "How do we deploy AI?" and began asking "Where can AI fundamentally reshape our competitiveness?"

Starting from business outcomes, the company focused on a small set of high-value bets that combine multiple AI use cases—faster product development, higher commercial productivity, and reduced unplanned customer downtime—to deliver outsized impact. The result? A more transformative agenda designed to create durable competitive advantages.

The lesson is to concentrate effort where AI can change the trajectory of the business. This requires future-back planning—designing from the AI-enabled enterprise you want to become, not from today’s pilots.

For leaders, this shift means treating AI as a core part of enterprise strategy rather than a parallel technology workstream. It needs disciplined capital allocation, executive accountability, and governance that ties AI directly to business outcomes.

Critically, it also requires measuring what matters. Successful scalers are significantly more likely than early explorers to embed AI key performance indicators (KPIs) into core business reporting—tracking enterprise value, not just activity. 

Figure 2

But waiting for the perfect measurement framework can slow momentum. The priority is tangible progress toward strategic objectives. Industry leaders will be those that move quickly and adapt as the technology evolves.

Modernize the workflow, not just the tools

Too often, companies bolt AI onto existing processes in pursuit of quick wins. The result is predictable: incremental efficiency gains constrained by legacy workflows, rather than breakthrough performance gains. Successful scalers take a different approach. They reimagine workflows alongside AI adoption—69% report strong process redesign capabilities vs. just 5% of early explorers.

Leading firms start with a clean sheet. They define the desired outcome, ask how AI could solve the problem end to end, and determine where human judgement remains essential. Only then do they redesign the workflow. This allows AI to operate beyond legacy models designed for a human-only world and deliver its full potential.

One leading pharma firm is pushing this further by redefining the role of human judgement. It asks a more practical question: Where is a human truly required to make a decision, and where is oversight enough?

This distinction—human-in-the-loop vs. human-on-the-loop—shapes how work gets redesigned. By encoding expert judgement, decision criteria, and risk tolerance into AI-enabled workflows, the organization can allocate human involvement more deliberately across the points where it adds the most value.

But reimaging the workflow is only the start. Execution requires the right operating model and incentives.

AI leaders are roughly five times more likely to deploy cross-functional agile pods, reflecting the reality that the talent required to reimagine workflows with AI spans multiple functions and a range of skill sets. At the center sits the product manager, bridging business, technical, and process domains; owning workflow redesign end to end; and holding accountability for value realization. Without that integrating role, even well-resourced efforts fragment back into silos.

Figure 3

Some of the people best positioned to reimagine a workflow are those who know it best today—and who are the most exposed to the change it will bring. Organizations that disincentivize bold thinking, even inadvertently, due to the workforce implications of those ideas will get incremental proposals, not transformative ones. The strongest companies align incentives to enterprise value creation. They minimize the uncertainty around job security and the familiar ways of working that leads teams to play it safe.

Even then, workflow redesign rarely goes according to plan. The real challenges often appear only in execution.

A pharma firm building an AI content production system offers a window into how this plays out in practice. The workflow was carefully mapped, built for physician-level personalization, and reflected how marketers and legal teams actually work. But only in production did the true granularity of the challenges emerge: gaps in system integration, unclear ownership, and breakdowns in handoffs.

These issues cut across systems and teams, meaning that no single function can resolve them alone. This is where C-suite sponsorship becomes critical. Leaders need the tolerance for deviation and authority to resolve issues across boundaries. Nearly half of successful scalers report institutionalized C-suite sponsorship for their AI initiatives, compared with just 4% of early explorers.

Workforce modernization across the organization

Redesigning the workflow changes the work—and the workforce required to deliver it. Large teams of functional specialists give way to smaller groups of talent operating across functional and technical domains. They compress long sequential cycles into faster, iterative ones.

This model is especially foreign to life sciences, an industry built on deep expertise and long development cycles. New roles, such as the AI product manager, are scarce and often need to be built proactively.

That makes HR a strategic partner from the start, helping identify talent—internally and externally—that can operate beyond strict job descriptions and functional boundaries. The data is clear: Successful scalers are nearly three times more likely to involve HR in workforce planning and upskilling, and 80% engage in future-back workforce planning vs. just 7% of explorers.

Figure 4

Some organizations are turning this into a structural commitment. Moderna, for instance, brought technology and HR together under a single chief people and digital technology officer to support more integrated workforce planning.

However, having the right roles and structure in place is not enough. Lasting value only comes when people adopt new ways of working.

AI is creating personal and professional uncertainty that few technologies have produced at this scale, and the human side of it—how people adapt, adopt, and sustain new ways of working—is the top-ranked barrier to value across all archetypes. Traditional change management was designed for discrete transitions with a defined end state. AI has neither.

Instead, AI demands an always-on change management model with visible sponsorship from the C-suite to the frontline manager. Leadership needs to participate openly, communication needs to be continuous, and training needs to evolve alongside the technology—reinforcing trust that the organization's AI ambition and its investment in people are moving together. Without that trust, neither adoption nor scale will follow.

Successful scalers are acting on this: 93% have implemented upskilling programs at meaningful levels, vs. 61% of early explorers.

What leading executives are doing differently

For life sciences leaders, the implication is clear. Scaling AI requires a different set of decisions—and a willingness to act on them now.

• Focus on a small number of transformative bets tied to enterprise priorities.  With a clear vision of the future state, build out a roadmap of use cases to get there.
• Fundamentally rethink how work gets done, including the operating model. Instead of bolting AI onto existing processes, redesign work from the outcome back with cross-functional agile pods, clear end-to-end ownership, and incentives aligned to enterprise value.
• Expect the unexpected. In a transformation of this magnitude with an ever-evolving technology, constraints and obstacles will surface that no plan can fully anticipate. Build tolerance for deviation, and step in to resolve issues across boundaries when they do.
• Plan the workforce future-back—and build toward it now. Redefine the roles and capabilities that AI-enabled workflows require, and proactively close gaps through workforce planning, reskilling, and early HR partnership.
• Lead the behavioral shift, not just the technology rollout. Strengthen trust and reshape incentives, decision rights, and leadership behaviors to enable new ways of working.

With the speed of AI development, the advantage will come from how quickly organizations adapt and act. Those that move decisively will pull ahead. Those that don’t will find the gap increasingly difficult to close.

About Mayfield Fund

Mayfield Fund is a leading venture capital firm that backs early-stage entrepreneurs building impactful technology companies.