Planning to scale or planning to fail? Organizing for scaled AI in pharma

From enhancing patient outcomes and experiences to plugging the world’s alarming doctor shortage, AI holds vast promise to transform healthcare. With drug companies already using AI to dramatically shrink development cycles, reduce research failure rates, improve customer (and patient) experiences and more, the pharmaceutical industry is no exception.
 

But companies striving to create lasting competitive advantage face a choice: Do they move swiftly, aiming for quick wins in a fast-paced market? Or do they take a more deliberate approach, focusing instead on long-term growth and stability?
 

How each company approaches this question will play a significant role in the ultimate success or failure of its AI initiatives.

Imagine two large pharma companies seeking competitive advantage by infusing artificial intelligence across their value chains. One lets a thousand flowers bloom, empowering local teams to rapidly build and implement use cases in search of quick wins. The other houses AI development, production and governance in a centralized center of excellence with an eye to sustainable expansion.
 

So, which company will finish ahead? Before answering this question, it’s crucial to understand how organizational archetypes for AI adoption shape the operating model, governance processes and speed and scale of an organization’s AI practices.
 

Today, nearly every pharma company begins its AI adoption journey by adhering to one of two organizational archetypes:

AI organizational archetype #1: The hub model. This model centralizes AI capabilities within a single enterprise-led team. Roughly 60% of organizations start here. Positives: Plentiful funding, expertise and leadership. Negatives: Initiatives may grow disconnected from business needs, as C-suite support is often required, but the top-down vision may become disconnected from on-the-ground business- or function-specific objectives.
 

AI organizational archetype #2: The spoke model. This model decentralizes AI efforts, empowering business units to lead their own initiatives. Roughly 40% of organizations start here. Positives: AI initiatives align with wider business needs. Negatives: Effort duplication and reliance on unit funding. Example: If multiple teams require a standard operating procedure co-pilot, letting each team develop its own solutions might be faster initially; however, a common scalable platform would be preferable in the long run.
 

But there’s a third organizational archetype for AI development, which is becoming more common because it leverages the strengths of both models:
 

AI organizational archetype #3: Hub-and-spoke model. This model combines centralized strategic oversight and scaling with decentralized innovation, allowing organizations to leverage collective expertise while maintaining agility. For this model, individual business units (“spokes”) identify, prioritize and develop tailored AI use cases, while a centralized center of excellence (“the hub”) provides strategic oversight, governance and reusable capabilities. While most companies begin their AI journeys organized around archetypes one or two, the hub-and-spoke model is more durable in the long run.

Following a hub-and-spoke model (Figure 2) allows pharma companies to centralize aspects of AI adoption that depend on ownership of the big picture while decentralizing those that benefit from proximity to end users. Use case teams (whether centralized or decentralized) progress through the AI development life cycle, centralized strategy, governance and capability assessment—which includes enterprise AI capability mapping and smart sequencing practices. This process allows organizations to generate a reusability cycle in which use case solutions remain durable, and therefore valuable, over time. Centralizing value measurement ensures consistency in monitoring business impact and effectiveness across the organization.

While the hub-and-spoke model offers many advantages over traditional ones and is optimal for most organizations, this approach comes with potential challenges organizations will have to plan for and mitigate against, including:
 

Managing the tension between control and enablement. While the hub is meant to streamline processes, it risks slowing things down. Organizations must carefully distribute decision-making power between the hub and the spokes to avoid the hub becoming a bottleneck.

Role clarity and governance. Companies must clearly delineate roles and responsibilities and institute robust governance frameworks to prevent overlap and ensure consistent compliance.
 

Culture and scalability. Overcoming resistance to change and ensuring that the organization can adapt and scale AI initiatives across multiple business units requires strategic planning.
 

Balanced performance metrics. Developing KPIs that align with both enterprisewide objectives and business-unit-specific goals is crucial for measuring success.

Accountability. While measuring outcomes is valuable, holding functions accountable is the bigger challenge. Implementing a no-code assistant that delivers productivity gains, for example, requires mechanisms to ensure proper use of the tool and value realization.

For pharma companies to realize the full value of their AI investments, scaling them across the enterprise is essential. Yet many organizations struggle to apply an integrated approach to their development and deployment. We suggest they begin with these steps:
 

Step 1: Develop a strategic roadmap for AI. Conduct a maturity assessment to gauge current AI capabilities, including data infrastructure, technology capabilities, talent proficiency and company culture vis a vis innovation, collaboration and adaptability. Using this baseline, companies should then develop a strategic roadmap with realistic milestones to ensure that AI projects align with business objectives.
 

Step 2: Establish an AI operating model. Establish an operating model that builds cross-functional collaboration frameworks for setting organizational guidelines on topics like risk management, responsible AI and use case prioritization. Once these guiding policies are in place, the cross-functional collaboration body can shift to testing and learning and building AI capabilities.
 

Step 3: Build for value realization. Demonstrate AI's value through tangible outcomes like cost savings, revenue growth and operational efficiencies. Educate users to address cognitive biases, mitigate resistance and build a culture of innovation that adopts and trusts AI. Embracing change and fostering a culture that values creativity, collaboration and continuous improvement will guide pharmaceutical companies toward becoming AI-driven innovation engines.
 

Step 4: Create and deploy a use case prioritization engine. To ensure that greenlit use cases align with a company’s strategic roadmap, it’s imperative for companies to develop and operationalize a use case prioritization framework (Figure 3). This requires:

• Strategic alignment, including defining a cohesive prioritization strategy aligned to overall business goals and then inviting employees to bring innovative ideas.
• A comprehensive intake mechanism, requiring detailed information on value, risk, strategic fit and buy-in for ideas. Set up baseline value measurements early to track impact and embed roles within business units to facilitate high-impact use case ideation.
• A structured decision-making methodology that involves a diverse stakeholder group for broader buy-in of AI initiatives.

AI’s potential to address emerging challenges in healthcare is clear, and the life sciences industry is no exception. As companies rapidly adopt AI, we’re starting to witness its transformative impact—especially in pharma. In fact, by scaling just seven common digital programs, the average top-10 pharma company stands to increase its operating revenue by as much as $2 billion over five years.
 

With AI's computational power significantly outpacing Moore's law, companies that fail to act quickly risk falling behind. This doesn’t mean that building a sustainable AI infrastructure can or should happen overnight. Sustainable AI infrastructure requires a strategic approach anchored to a clear vision, robust governance and a focus on value realization.

While the hub-and-spoke model is particularly effective, it may take time for organizations to get there given existing strategic priorities, organizational structures and funding mechanisms. Companies can still focus on maturing capabilities within a hub or spoke model as they gradually transition to a more integrated approach. Those that act quickly and get it right will unlock substantial cost savings, improved efficiencies and enhanced patient outcomes, ensuring they position themselves at the forefront of the AI-driven revolution.