AI capabilities are advancing rapidly, but are pharma organizations adapting fast enough to realize their full potential? As companies explore new technologies, the challenge is increasingly about turning AI investments into measurable business outcomes.
At the Indegene Digital Summit 2026, Agam Upadhyay, Chief Digital and Information Officer at Vertex Pharmaceuticals, shared his perspective on what it takes to bridge this gap. The discussion offered practical lessons on AI strategy for CIOs and moving beyond isolated AI initiatives to rethink how decisions are made, work gets done, and value is realized.
The Challenge: Why AI Progress Does Not Always Translate Into Business Value
Traditionally, technology teams have focused on delivering solutions against strategies defined by business functions. AI is challenging this model, creating opportunities to rethink how organizations approach clinical development, molecular design, commercial strategy, and operations. Yet, the path from AI adoption to business value remains difficult. Organizations may accumulate large portfolios of disconnected use cases, continue experimenting without clear priorities, or achieve incremental productivity gains without a defined approach to realizing their value.
Saving 10% or 20% of someone's time is valuable, but what happens to that capacity? How does it translate into faster execution, improved capabilities, or more efficient spending?
Addressing these questions requires a more deliberate approach to AI transformation, beginning with how organizations set priorities and extending to how they redesign and govern work.
1. Move Technology Leadership Upstream into Business Strategy
AI transformation needs to start where business priorities are defined. The traditional sequence of business strategy followed by technology implementation can limit opportunities to rethink established processes.
Technology leaders need to work alongside commercial, R&D, manufacturing, and other functional leaders to identify where AI can change outcomes. This could mean bringing AI into molecular design earlier, reconsidering clinical trial workflows, or exploring quality-by-exception approaches in operations.
The shift also requires stronger business ownership. Technology teams bring the capabilities and expertise, while business leaders help identify the problems worth solving and take ownership of the outcomes.
2. Build a Trusted Data and AI Foundation Before Scaling
Once AI becomes part of business strategy, the next challenge is establishing a reliable foundation to support it. Well-managed data platforms, curated datasets, consistent information, and reliable data pipelines are essential. The discussion highlighted the importance of investing in these capabilities before pursuing AI transformation at scale, so that initiatives can work from a common source of truth.
Equally important is providing employees with secure, compliant access to AI. Enterprise-wide platforms, supported by employee education, can help make these capabilities accessible while maintaining a consistent environment for their use.
This combination of data readiness and broad access creates the foundation for more complex business applications. The priority is to build capabilities that different functions can use and build upon, rather than approaching every AI opportunity as a separate initiative.
3. Replace Expanding Use-Case Lists with Focused, Value-Led Investments
With the foundation in place, prioritization becomes the next test.
Agam pointed out that rejecting weak ideas is relatively straightforward. The more difficult decisions arise when organizations must choose between several promising opportunities. A disciplined approach starts with business leaders identifying a focused set of high-value priorities. Each initiative should have a clearly defined business need, expected investment, and measurable outcome, whether that involves improving speed, efficiency, capabilities, or spending.
A value realization framework can help evaluate these priorities and establish how the expected benefits will be measured and captured. This is particularly important when productivity improvements are spread across multiple activities and teams.
The approach also calls for greater discipline around experimentation. Rather than maintaining a large collection of disconnected pilots, organizations need to concentrate resources on initiatives with a clear path to production and business value.
4. Challenge Existing Processes Instead of Automating Individual Tasks
Prioritizing the right opportunities is only part of the equation. Capturing their full potential often requires rethinking the underlying process.
One of the strongest examples from the session came from clinical database development. According to Agam, the process typically takes 12–14 weeks across the industry. Vertex had already shortened its timeline to eight or nine weeks when its CEO challenged the team to complete it in just one day.
That ambitious target pushed the team to rethink the entire workflow and introduce an AI agent framework, rather than simply automate existing steps. The result was a reduction to five days, with two clinical studies already using the redesigned process. Although the original one-day target was not achieved, the improvement demonstrated what becomes possible when leaders challenge established assumptions about how work should be done.
5. Enable Business-Led Execution with Clear Enterprise Guardrails
As AI initiatives move into production across functions, another challenge emerges: how to support decentralized execution without creating fragmented technology, inconsistent governance, or duplicated investments.
One approach discussed was establishing a lean central transformation office that provides oversight across three critical areas: technology architecture, investment and value realization, and continuous learning across implementation waves.
Business teams can lead execution, while the central structure maintains consistency and helps lessons from one initiative inform others. Responsible AI governance also needs dedicated attention. A cross-functional council involving legal, privacy, HR, cybersecurity, and architecture can review AI models and capabilities. Clear policies, supported by usage monitoring, help organizations maintain visibility into how AI is deployed and ensure that security and compliance requirements are addressed.
6. Make Employees Part of the Transformation, Not Just Its End Users
Even with the right technology, priorities, and governance, AI adoption depends on how employees engage with new ways of working.
The discussion challenged the effectiveness of relying primarily on training, emails, and internal communication campaigns to drive change. Employees need opportunities to explore what AI can enable and contribute ideas based on the challenges they encounter in their work.
A company-wide hackathon organized by Vertex with strategic partner Snowflake illustrated this approach. It encouraged employees to participate directly, generated strong engagement, and surfaced business ideas that subsequently became part of the company's prioritized AI initiatives. The example also highlighted the role strategic technology partners can play in helping organizations explore new possibilities and encourage innovation.
Looking Ahead: Organizational Readiness Will Shape AI's Impact
As AI agents increasingly work alongside people, pharma leaders will need to rethink workflows, responsibilities, and operating models while retaining human expertise and judgment. Closing this gap will require a connected approach to business strategy, data readiness, focused investment, process redesign, governance, and employee engagement. The central lesson is clear: realizing AI's value depends as much on an organization's ability to change how it works as on the technology it adopts.