Life Sciences Practice

Scaling gen AI in the

life sciences industry
Gen AI pilots have shown promise, but for the technology to deliver
transformational business value in the life sciences industry,
organizations need to rethink how they scale it.
by Chaitanya Adabala Viswa, Dandi Zhu, Delphine Zurkiya, and Joachim Bleys

January 2025
Back in July 2023, researchers at the McKinsey Global Institute estimated that gen AI could
unlock between $60 billion and $110 billion a year in economic value for the pharmaceutical and
medical products industries, boosting productivity and innovation in domains across the
industry’s value chain—from the way new treatments are discovered to how they are marketed
and administered by physicians. Six months later, McKinsey experts dug deeper into those
numbers, uncovering more than 20 use cases with the greatest potential for near-term impact.

Now, with gen AI use cases proliferating across the business community, we decided to find out
how much progress life science organizations have made in capturing this value. In late summer
2024, we surveyed more than 100 pharma and medtech leaders responsible for driving their
organizations’ gen AI efforts. All respondents report having experimented with gen AI, and 32
percent say they have taken steps to scale the technology. But only 5 percent say they have
realized gen AI as a competitive differentiator that generates consistent and significant financial
value (Exhibit 1). Nonetheless, companies remain optimistic about gen AI, with more than two-
thirds of respondents saying they plan to significantly increase investment in the technology
(Exhibit 2).

Exhibit 1

Scaling gen AI in the life sciences industry 2

Exhibit 2

Why do so many life science organizations struggle to realize results from their gen AI
deployments? And what are the minority of top performers doing differently? This article reveals
the most common pitfalls life science companies are facing—and offers solutions that can help
organizations move from pilot purgatory to driving real business value at scale.

The key challenges to scaling gen AI in life sciences

Based on our survey and our experience, we have identified five key areas that pose challenges
for life science companies attempting to realize company-wide value from gen AI: gen AI
strategy, talent planning, operating model and governance structure, change management, and
risk (Exhibit 3).

Scaling gen AI in the life sciences industry 3

Exhibit 3

Challenge 1: Ambiguous, shortsighted, or nonexistent enterprise gen AI strategy

About 75 percent of respondents say that their organizations lack a comprehensive vision for
gen AI or an intentionally designed, strategic road map with clearly defined success measures
linked to business priorities. Instead, they tend to proceed in a decentralized manner, use case
by use case. This instinct to capture short-term value through experimentation, coupled with the
federated/function-led structure of many life science organizations, explains many of the
challenges organizations encounter when it comes to scaling.

McKinsey research has found that digital transformations seldom succeed unless C-suite
leaders are aligned around a business-led road map. Without an intentional strategic posture
toward gen AI—whether a top-down mandate or a coordinated enterprise road map driven by a
center of excellence—individual business units are left to navigate the ever-evolving technology
landscape on their own, pursuing a multitude of new use case ideas that, no matter how
compelling, often fail to add up to a strategy that delivers actual value.

Scaling gen AI in the life sciences industry 4

Challenge 2: Lack of talent planning and upskilling
At most life science companies, the existing pool of tech talent presents a traditional tool kit for
IT, data science, and product development. Unfortunately, traditional approaches to tech talent
are unable to deliver the quality and performance of enterprise-grade solutions needed for gen
AI, for example, agent-based architecture, model validation, large language model (LLM)
operations, and the fine-tuning of models. But only 6 percent of survey respondents report
having conducted a skills-based talent assessment to determine how to evolve their talent
strategy into one that considers gen AI priorities.

Prompt engineering has emerged as a key gap, especially for more complex gen AI applications.
One life science company, for example, was attempting to use gen AI to draft regulatory
documents, only to discover that prompt engineers required a unique combination of regulatory
domain knowledge and engineering rigor to craft scalable prompts that generate submission-
ready output—a specialized necessity that made the role especially challenging to fill.

Challenge 3: Loosely defined operating model and governance

One common challenge leaders face is creating the right operating model for gen AI
transformation, often choosing between one of two extremes. At one end of the spectrum is a
highly decentralized approach, in which the organization simultaneously launches multiple use
case pilots. While this allows companies to move fast, it also leads to quality, cost, and
sustainability challenges and creates operational silos that inhibit the sharing of knowledge and
the ability to capture cost synergies. At the opposite end is a top-down approach, with
centralized decision-making and a phased rollout of use cases. This approach can be slow and
often frustrating, destroying momentum.

One company swung between the two. It began its gen AI efforts by launching 1,500 different
use cases. When that proved unwieldy, company leaders imposed a top-down governance
structure that led to a different set of issues, constricting the innovation pipeline with projects
requiring an arduous approval process that stretched some two to three months.

Challenge 4: Underestimating the process rewiring required to drive scale

To succeed with gen AI, companies must integrate the technology across complex workflows to
promote adoption and impact—a reality that highlights the need for effective change
management. McKinsey has found that 70 percent of digital transformations fail not because of
technical issues but because leaders ignored the importance of managing change. In fact, for
every $1 spent on technology, $5 is required for change management to successfully drive
capability building, adoption, buy-in, and value capture over time.

One company launched a center of excellence function to initiate a broad gen AI platform for a
range of use cases but failed to communicate a compelling change story to accompany those
initiatives. That failure, coupled with the lack of holistic, end-to-end planning and thinking,
resulted in a collection of gen AI tools that almost no one ended up using.

Scaling gen AI in the life sciences industry 5

Challenge 5: Inadequate understanding of risk
Gen AI introduces unique risks, from hallucinations and accuracy to bias and intellectual
property protection. But 35 percent of survey respondents report that they spend fewer than
ten hours with their risk counterparts, limiting the degree of collaboration with these crucial
functions. This dynamic needs to evolve to scale gen AI. Successful scaling requires business
leaders, technology teams, and risk management professionals to communicate from the outset;
the absence of such collaboration can lead to issues being raised late in the game, when they
are much more difficult to fix, or a lack of adherence to the risk and compliance guardrails that
are critical to building trust in the organization.

One company, for example, spent several months developing an external-facing gen AI solution,
only to be forced to withdraw the launch due to a lack of alignment with its digital, medical, and
legal teams—which raised significant risk issues after the tool had been developed. This resulted
in a severe setback for the gen AI team’s agenda, morale, and momentum.

The solution: A five-point plan to realize value from gen AI

Successfully scaling gen AI and capturing its value potential requires more than just a
technological rollout. An effective gen AI strategy is fundamentally different from traditional tech
projects. Given the rapid pace of innovation, a gen AI strategy must be dynamic, scenario driven,
and focused on how to engage with the broader ecosystem. Scaling gen AI involves
comprehensive change across the organization, encompassing strategy, talent, governance,
and risk management.

Based on our experience, we have identified five key strategies to move from gen AI use cases
to enterprise-wide adoption. These actions ensure that organizations not only experiment with
the technology but also fully integrate it into their operations to drive measurable business
value.

— Adopt a domain-driven approach. Successful AI strategy cannot be based on a slew of

disconnected use cases, which often leads to fragmented efforts and missed opportunities.
Instead, the focus must shift to domain-driven transformations, where gen AI is applied to
fundamentally reshape critical areas of the business, such as the commercial, medical, or
R&D domains. Thirty-eight percent of the life science organizations surveyed cite research
as their leading strategic priority in their gen AI journey, followed by the commercial domain,
at 28 percent (Exhibit 4).

This domain-driven approach ensures that gen AI isn’t just another tech solution but a core
enabler of business transformation. Rather than focusing on technology for technology’s
sake, organizations that prioritize domain transformations are better positioned to capture

Scaling gen AI in the life sciences industry 6

 the full value of AI. Crucially, there is no such thing as a stand-alone gen AI strategy. The real
focus should be on deploying gen AI to support broader business objectives, drive strategic
goals, and create differentiation in the market. Organizations that view the technology
through this business-first lens have found greater success in scaling AI initiatives.

Exhibit 4

Scaling gen AI in the life sciences industry 7

— AI transformation encompasses more than just tech. Scaling gen AI isn’t simply a matter of
implementing a new technology; it’s about rewiring the organization’s operating model and
culture to support new AI-driven ways of working. This extends to talent strategies: the
workforce must evolve beyond traditional IT data science roles to include new skills—AI
engineering, large language model fine-tuning, and business translation—to bridge the gap
between technical execution and business value capture. Without a comprehensive talent
realignment, organizations will be less successful in scaling their gen AI efforts. Further, gen
AI implementation needs to drive measurable value. This requires a clear up-front
agreement on how value will be captured, say, through acceleration of time to market,
productivity increase, or improved probability of success.

One life sciences company, for instance, launched an enterprise talent upskilling and planning
program, with targeted initiatives for business and technical roles. The program also
introduced dedicated gen-AI-focused leadership roles in critical functions to drive sustained
organizational change. With the appropriate talent—and leadership—in place, the company’s
gen AI initiatives proceeded much more smoothly than they would have otherwise.

— Adopt an ecosystem approach. In the rapidly evolving AI ecosystem, an externally focused

partnership strategy is critical. Given the speed at which AI technologies and methodologies
are advancing, life sciences organizations should consider cultivating a network of low-cost,
high-optionality partnerships. These partnerships can provide flexibility and give
organizations the ability to quickly pivot and seize opportunities as they arise. Organizations
should also establish clear “triggers” that indicate when it’s time to move from exploratory
partnerships to larger strategic bets. This ensures that the business remains agile and can
scale up or shift its AI investments based on real-time insights and market movements.

Engaging with the broader ecosystem—including academia, tech, and venture capital—is also
essential to staying on top of the latest developments. Relying solely on internal capabilities
is no longer enough to stay competitive in AI. A dynamic, externally focused lens ensures
that companies stay ahead of the curve and capture the full value of gen AI innovations.

— Deploy a platform-driven approach from the outset. A platform-driven approach is key to

ensuring that gen AI initiatives are scalable, sustainable, and reusable across various
business domains. A scalable AI platform allows organizations to standardize infrastructure,
data pipelines, and development processes, ensuring that each new use case builds on the
previous one. This can also help reduce duplication of effort, encourage collaboration across
business units, and foster consistency in AI performance across the organization. Moreover,

Scaling gen AI in the life sciences industry 8

 a platform-driven approach ensures that AI models are not developed in isolation but are
integrated into a unified framework, allowing them to be adapted and reused across various
business domains. This not only reduces costs but also accelerates time to value, as insights
from one domain can be applied to another.

One life sciences company found success by adhering to a mantra: “Slow down to speed up.”
The company spent three months defining a detailed blueprint for insights and document
platforms. This enabled the reuse of components within each platform, enabling rapid
scaling across use cases.

— Embed risk management in the full product development life cycle. One of the common
mistakes organizations make with gen AI is treating risk management as an afterthought or
as an obstacle to innovation. In fact, risk management must be embedded throughout the
entire AI product life cycle. Gen AI introduces unique risks—such as hallucinations, bias, data
security, and intellectual property issues—which require careful oversight.

To ensure these risks are managed effectively, business leaders and risk and compliance
functions should collaborate regularly. Organizations should establish clear governance
frameworks early on and ensure that ethical guidelines are in place to address concerns about
AI fairness, transparency, and accountability.

Given the high regulatory requirements in life sciences, organizations should place greater
emphasis on risk management. One organization proactively identified the guardrails necessary
to address evolving regulations (for example, the EU AI Act) and technology limitations (for
example, the probabilistic nature of models). The organization established clear, responsible AI
requirements, including mandatory observability, validation protocols, and human-in-the-loop
guidelines, which were defined prior to the start of product development.

What a holistic transformation can look like

What does a successful gen AI initiative look like? Consider one life sciences company that
recognized the gen AI opportunity early and embarked on a holistic transformation across
domains. Company leaders convened a C-level task force to steer the overall gen AI strategy,
set up governing bodies across the R&D, commercial, medical, and operations domains, and
asked each domain to prioritize one use case with high-value potential for C-level sponsorship.

The company then ran proofs of concept with an eye toward scaling, using its early experiences
to organize reusable components into domain-specific platforms. The technology and business

Scaling gen AI in the life sciences industry 9

teams partnered from the outset, ensuring that all gen AI solutions addressed priority business
needs and helped drive the process changes needed to spark adoption and deliver value.

In the meantime, the company engaged ecosystem partners to bring in learnings and assets
from across the life sciences industry and beyond and built stage gates to focus resources on
partnered solutions that were ready to scale across therapeutic areas and geographies.

Leaders shaped a compelling change story focused on how gen AI solutions were intended to
augment rather than replace employees, for example, by helping them deal with increasing
workloads, and used change management teams to help drive a successful rollout. They
provided white-glove support for initial users and deployed these early adopters as change
ambassadors to build bottom-up momentum. Impact metrics were defined, tracked, and
reviewed at regular governance meetings to ensure gen AI initiatives remained on track to scale
and deliver business impact.

This kind of experience does not have to be an outlier. Leaders of life science organizations
should understand that capturing the potentially transformative value of gen AI requires more
than experimentation and individual use case deployment. It demands strategic integration into
the organizational fabric. In the next chapter of the gen AI story, organizations should take an
intentional approach to driving alignment with business strategy, scalability, and sustainability.
This pivotal moment is an opportunity for life sciences leaders to lead transformative change,
revolutionizing drug discovery and patient care, as well as driving meaningful bottom-line
results.

Chaitanya Adabala Viswa is a partner in McKinsey’s Boston office, where Delphine Zurkiya is a senior partner;
Dandi Zhu is a partner in the New York office; and Joachim Bleys is a senior partner in the Carolinas office.

The authors wish to thank Abhi Mukherjee, Lionel Jin, Natalia Dorogi, Nitisha Sharma, and Vasu Macherla for their
contributions to this article.

This article was edited by Larry Kanter, a senior editor in the New York office.

Copyright © 2025 McKinsey & Company. All rights reserved.

Scaling gen AI in the life sciences industry 10