In a recent ZS survey of technology executives at life sciences companies, executives agreed AI (and specifically generative AI) will be the top driver of their companies’ priorities in the year ahead—outpacing digital innovation, rising costs, and drug pricing and reimbursement constraints.
Many companies’ early forays into generative AI involve experimenting with use cases that drive productivity and efficiency—a natural choice as a low-risk play to gain confidence and trust in this new modality. For this type of use case, companies are using generative AI to solve problems they already know how to solve using legacy techniques, but gen AI helps them do it faster and with the added security of using humans to “check the work.”
However, most companies today are stuck in a transitional phase with generative AI where enthusiasm outstrips impact. Here’s what we’re hearing from leaders in life sciences:
- Companies are investing in disjointed use cases that create productivity gains in isolated pockets but fail to drive the business objectives that executives are targeting.
- Early enthusiasm for new tools is fizzling in the face of poor experiences, oversaturation and tenuous results. In addition, rank-and-file employees feel alienated from their companies’ generative AI efforts because leaders haven’t sufficiently articulated their vision for it.
To begin truly scaling AI for value, companies need to shift focus beyond seeking productivity and efficiency gains toward delivering fundamental shifts in what they do and how they do it. This requires a clear vision aligned with business goals and the leadership support to develop whole chains of use cases that make that vision a reality.
Generative AI allows for new links in this chain, but classical AI and automation also must play a role. Companies with a value mindset will judiciously pair classical AI with generative AI to boost productivity, produce deeper insights and drive better decisions. Imagine predicting patient behavior (a classical AI problem), mining unstructured data to better understand patient needs and motivations (generative AI for deeper insights), quickly synthesizing vast troves of business intelligence (generative AI for productivity) and then generating content and simulating patient reactions to them (generative AI for better decisions) to create better engagement.
For generative AI solutions, think beyond productivity gains
Classical AI excels at making predictions based on patterns in structured data and has shown promise in many areas of healthcare, such as clinical decision support. However, these models depend on having access to large volumes of structured data. Generative AI, on the other hand, can mine unstructured data (which makes up some 80% of all healthcare data) and enables users to do things they couldn’t before, such as pulling insights from far-flung, unstructured data at scale and generating new content.
Generative AI use cases fall into three broad categories, representing a clear hierarchy of organizational value. They can be used as:
Productivity enhancers. These use cases help companies do the work they’re already doing faster. Think automating repetitive tasks, performing analytics, summarizing reports and locating data.
Maturity: high | Business value: low
Insights generators. These use cases fall under the general umbrella of automation, but they’re potentially much more valuable because they draw insights from data that traditional analytical techniques can’t. Think mining unstructured data to generate deeper customer insights, synthesizing insights from multimodal data, and detecting patterns and anomalies in unrelated data sets.
Maturity: medium | Business value: medium
Enablers of better decisions. These use cases aim to support decision-making in new ways and are key to reengineering business processes. Think creating custom marketing content tailored to an individual customer, simulating a physician’s reactions to alternate messages and then identifying the optimal next action.
Maturity: low | Business value: high
So far, companies have focused investment and attention on enabling generative AI pilots that supercharge productivity. In our own business, for example, analysts equipped with operational analytics co-pilots have experienced productivity gains of up to 50%. It’s easy to understand why companies start here, as these use cases are low risk, comparatively easy to pull off and offer a way to get people comfortable with the technology. But tools that drive productivity gains alone will no longer be differentiating as more companies build and deploy these capabilities, either with homegrown tools or third-party integrations such as Microsoft Copilot.
True competitive advantage will accrue to those companies that combine use cases from all three categories above and pair them with their legacy AI to fundamentally reimagine wholescale business processes.
FIGURE 1: Scaling AI for future value
To scale generative AI, think interconnected chains—not collections of use cases
Experimenting with a variety of generative AI use cases is important and should not be discounted. However, focusing investments into dozens of use cases without tying them to the digital programs already proven to demonstrate value will leave companies in a state of paralysis induced by “death from a thousand use cases.”
To get unstuck and scale AI for value, leaders should start with the business outcomes they want to drive and then work backward to identify each interconnected use case that contributes to those outcomes. Each use case must then be individually engineered with a view of the big picture.
While generative AI sits at the top of the corporate agenda for good reason, companies can’t afford to ignore the other AI programs they’ve been developing during the past decade or more. It’s still best suited to making predictions based on historical patterns. The key to building valuable, sustainable AI solutions is in how companies bring classical and generative AI together to transform business processes across the entire pharma value chain.
Generative AI in practice: Clinical development
Patient enrollment challenges plague roughly 85% of clinical trials, with overly complex trial protocols, operational inefficiencies and poor data management (among other challenges) extending trial durations and exploding development budgets. Generative AI, especially when paired with existing AI programs, can be used to improve trial plans, design key statistical measures and optimize protocols—leading to faster trials, higher data quality, enhanced patient engagement and improved decision-making.
Generative AI in practice: Commercial pharma
Many pharmaceutical companies already have mature approaches for leveraging classical AI systems to optimize omnichannel execution for healthcare provider (HCP) engagement. Figure 3 illustrates how building generative AI capabilities on top of existing ones can create a more fundamental shift in HCP engagement by moving further upstream.
Generative AI practice: Supply chain and manufacturing
Resilience, agility and sustainability are the pillars of a stronger and more advanced supply chain and a key source of future competitive advantage. Figure 4 illustrates how a company might pair generative AI with AI to build real-time resilience agents using up-to-the-second external data.
How to start building competitive advantage at scale with generative AI
Despite being foundational to companies’ quests to generate returns on their AI investments, fewer than half of life sciences executives we surveyed say they have a vision for bringing together classical and generative AI. Here’s a formula for moving beyond the productivity- and efficiency-enhancing phase of generative AI development and into one that leverages the best of AI to create enduring business value.
- Articulate a vision that’s tied to business objectives. Generative AI shouldn’t fundamentally change a company’s priorities, so assessing generative AI investment areas should start with first principles. Ask: Where can I use generative AI to double down on my business objectives? Pick two or three fundamental shifts that are needed to achieve these objectives and then empower dedicated teams to execute against the vision.
- Think three-in-a-box teams with AI, tech and domain. AI doesn’t work in a vacuum. Teams implementing AI solutions need an internal technology partner to ensure they’re building durable capabilities, and they need an internal domain partner to ensure the business context is brought to bear and end users adopt solutions. When AI, tech and domain come together, fundamental transformation is possible. This three-in-a-box team is responsible for mapping out and deploying durable assets and capabilities in the right sequence to bring the string of connected use cases to life for each fundamental shift leaders have prioritized.
- Measure outcomes and celebrate successes. Generative AI initiatives should be run like all other digital transformation initiatives, with value planned from the start. Leaders should measure success not only by adoption rates but also by the outcomes they drive. From the executive level down, everyone should be accountable for one or two measures related to changing behavior and driving metrics tied to tangible business value.
When leaders get this formula right, they create a rare type of virtuous cycle: a business transformation that’s driven from the top but nurtured and sustained by passion and energy from below.
Add insights to your inbox
We’ll send you content you’ll want to read – and put to use.
