AI Is Not a Technology Problem — It’s an Operating Model Transformation
Boards are asking about it. Executives are funding it. Teams are experimenting with it. Yet despite this momentum, many organizations are struggling to translate AI into meaningful, scalable value.
And yet, across the industry, a consistent pattern is emerging: many organizations are approaching AI as a technology initiative. And that’s a mistake.
The Technology Trap
In many organizations, AI efforts follow a familiar path:
evaluate tools
run pilot programs
test use cases
deploy isolated solutions
This approach feels logical. It mirrors how many prior technologies were adopted.
But AI behaves fundamentally differently from prior technologies — in its speed of evolution, breadth of impact, non-deterministic nature, and rapid rate of obsolescence (among other reasons). AI does not simply support existing processes. It changes how work is performed, how decisions are made, and how value is created. It affects not just systems, but the structure and behavior of the institution itself.
Why does this distinction matter? Because organizations that focus primarily on tools and pilots often experience poor outcomes with AI:
fragmented initiatives
higher error rates (e.g., hallucinations)
limited scalability
unclear value realization
growing complexity and risk
The issue is not the technology. The issue is how the organization is aligned to use it.
AI as a Capability Transformation Problem
The organizations that are beginning to realize meaningful value from AI are not treating it as a technology layer. They are treating it as a capability transformation problem. This means asking a fundamentally different set of questions.
What data foundations are required to support those decisions?
How should workflows be redesigned to incorporate AI outputs?
What governance and quality systems are needed to ensure reliability and trust?
What skills and roles are required to operate in this new environment?
How should decisions be made differently in an AI-enabled organization?
AI does not succeed because a model works. AI succeeds when an institution is able to consistently and reliably integrate it into operations.
A Pattern That Has Repeated Across the Industry
Though AI is unique, healthcare and life sciences industry veterans may see some similarities with prior technology-driven transformations.
Combinatorial Chemistry & High Throughput Discovery
One of the earliest transformations in modern drug development came with the rise of combinatorial chemistry and high-throughput screening. The promise was compelling: dramatically expand the number of compounds that could be generated and tested, accelerating the discovery of viable therapeutic candidates.
But success did not come simply from introducing new laboratory technologies. It required fundamental changes to how research organizations operated:
new experimental designs and screening methodologies
integration of automation technologies with laboratory workflows
development of data systems capable of managing and interpreting vastly larger volumes of experimental results
closer coordination between chemistry, biology, and computational teams
Organizations that treated combinatorial chemistry as a discrete scientific innovation often struggled to realize its full value.
Even successful adopters did not fully realize the promised value. While combinatorial chemistry expanded the scale of compound generation and screening, it did not proportionally accelerate drug discovery. In many cases, the limiting factors were not the technologies themselves, but the surrounding systems required to translate scale into value. Biological understanding did not advance at the same rate as compound generation, data environments struggled to manage and interpret the resulting volume of information, and organizational models remained fragmented across chemistry, biology, and computational disciplines. As a result, organizations were able to generate more data and more candidate compounds, but not necessarily better outcomes.
The lesson was clear: scaling one component of the system without redesigning the broader operating model limits the impact of even the most powerful innovations.
Digitizing Clinical Research
In the early transition from paper-based trials to electronic data capture, success did not come from simply introducing new software. It required new processes; new data standards; new operating models; and cross-functional coordination across scientific, technical, and operational teams. The organizations that treated it as a system implementation struggled. Those that treated it as an institutional shift succeeded.
Health Analytics
As the healthcare industry adopted electronic medical records, their data became computable. So we saw the emergence of investments focused in data and analytics platforms supporting hospital operations, value-based care programs, and more. But analytics platforms alone did not produce impact. Real transformation required:
building dedicated analytics capabilities that were adept at harnessing this data
aligning clinical and operational leadership on effective measurement strategies
adjusting care pathways, processes, and EMR functionality to leverage the insights
redefining decision-making processes based on the new insights available
establishing governance and prioritization frameworks protecting patients and the judgment of clinicians.
Analytics became valuable not because it existed, but because it was institutionalized.
The AI Era
AI is now following the same pattern, but at far greater speed and scale. In areas such as clinical research, real-world evidence, regulatory processes, and healthcare delivery, AI is already reshaping how work is performed, how quickly it can be performed, and what level of precision is possible. In CROs, for example, AI is shifting value away from labor-based models toward AI-enabled capabilities, fundamentally changing how services are delivered and evaluated.
At the same time, organizations are struggling with competing pressures: speed vs. governance, innovation vs. quality, and adoption vs. control. Regulators are also signaling expectations that reinforce this shift — emphasizing data governance, lifecycle management, and multidisciplinary expertise as core requirements for AI-enabled systems. These are not technology challenges; they are institutional challenges.
What It Actually Takes to Scale AI
Organizations that succeed with AI tend to converge on a common set of capabilities.
Clear Strategic Alignment. AI initiatives are directly tied to measurable business and clinical outcomes, not isolated experimentation.
Data as a Managed Asset. Data is treated as a product — governed, curated, and designed for reuse — rather than an output of individual systems.
Redesigned Workflows. Processes are re-engineered to incorporate AI outputs into real decision-making, rather than layered on top of existing workflows.
Governance and Quality Systems. AI is managed within structured frameworks that ensure reliability, compliance, and trust.
Integrated Operating Models. Cross-functional teams align business, scientific, and technical perspectives to deliver and sustain capabilities.
Continuous Adaptation. Organizations are designed to evolve as technologies change, rather than relying on static roadmaps.
Moving Beyond the Pilot Phase
Many organizations are still focused on identifying the “right” AI use cases. But that is not the limiting factor. The limiting factor is the ability to build the institutional capabilities required to scale those use cases.
Pilot programs are useful for learning. They are not sufficient for transformation. The organizations that move ahead will be those that shift their focus from “Where can we use AI?” to “What must we become to operate effectively in an AI-enabled world?” The real opportunity is not simply to do the same work faster. It is to transform the fundamentals:
redesign clinical research
improve the precision of care
accelerate therapeutic development
fundamentally change how organizations operate and the associated costs
The organizations that succeed will not be those that adopt AI tools the fastest. They will be those that build the capabilities required to make AI a core part of how they function.
AI is not a technology strategy issue. It is a transformation of how institutions create value. And the organizations that recognize this — and build accordingly — will define the next era of healthcare and life sciences.
