Fast 5: Can pharma-AI mega-deals transform commercial ops?

1. What’s happening

This year has seen a flurry of mega-deals involving large drugmakers and big AI players designed to accelerate drug discovery and other functions.

• Last Wednesday, Bristol Myers Squib said it will deploy Anthropic’s Claude across various company functions, including commercial.
• Google Cloud last month said it signed a deal with Merck, in which the latter committed up to $1 billion to deploy Google’s AI platform across its R&D, manufacturing, and commercialization units.
• Also in April, Novo Nordisk announced a partnership with OpenAI to integrate the ChatGPT maker’s AI capabilities “from drug discovery to commercial operations.”
• Nvidia and Eli Lilly inked another one in January to create a research lab in San Francisco as well as to explore AI’s application in commercial ops, the latest in a relationship which also spawned an Nvidia-powered supercomputer.
• More recently, Anthropic purchased AI biotech startup Coefficient Bio for $400 million and announced that Novartis’ CEO had joined its board.
• Meanwhile, pharma continues to partner with TechBio firms — last month Lilly signed a deal worth up to $2.25 billion with Profluent and in March entered a pact with Insilico Medicine worth up to $2.75 billion, while Pfizer did a deal with AI research firm Boltz, PBC, in January.

2. What’s the big deal

Pharma has invested billions in AI infrastructure, and cycled through multiple partners, in search of an AI sherpa to help embed the technology deeper into R&D and other areas of its business. While many have been running agentic AI pilots, the recent collaborations show they’re serious about moving from prototype into production environments.

• Life science companies’ investments in AI drug-discovery platforms have yet to yield a novel compound or lead to an approved product.
• In contrast to those smaller, siloed discovery efforts with the “AI-first” biotechs, dealing directly with the AI power brokers is a broader approach that could set companies up well to implement AI across their product lifecycle.
• Yet, as AI use expands beyond R&D, a larger question has arisen: To what extent may these pairings help pharma make fuller use of advanced analytics on the commercial side of their businesses, as well?

3. What’s at stake

The industry at large doesn’t want to get left behind by the AI boom. GLP-1 leaders Lilly and Novo see it as part of a plan to build non-obesity pipelines for the future, and Merck cites "one of the most significant launch periods in our company’s history.”

• Among other drivers, as regulators work to collapse drug review cycles, such as moving to single-arm trials and streamlining preclinical testing, AI’s exceptional speed and accuracy can help companies start preparing for commercialization a lot sooner.
• Then there’s the promise of embracing AI-operated commercial workflows which, studies suggest, can lead to significant competitive advantage in life sciences.
• A 2025 analysis cited by McKinsey showed that 75% to 85% of pharma workflows contain tasks that could be enhanced or automated by AI agents — from marketing and market access to sales and medical affairs — potentially freeing up 25% to 40% of an organization’s capacity.
• A 2025 IQVIA study, for instance, suggests a 15% to 25% lift in scientific engagement and follow-through when sales reps and MSLs access real-time copilots generating compliant next steps tailored to each HCP’s scientific interest and intent.
• But in many organizations, AI progress falls into a grey area, between deploying at scale and still trying to make it out of proof-of-concept stage.
• Only 22% of life sciences leaders self-report successfully scaling AI, with only 9% reporting significant returns, according to a 2026 Deloitte survey. And that’s just among those that are trying to do so — another 2025 McKinsey study found pharma and medical products companies ranked last among a dozen sectors in terms of using AI for various business functions.

4. What’s it mean for commercial execs

While pharma is still in the early stages of deriving enterprise value from the technology, industry is taking two big steps from theory into practice, says Deepak Mistry, a consultant at Beghou.

• First, AI is now being evaluated by measurable economic lift. The CTO, CIO, or head of data at pharma/life science companies has become accountable for showing value on both the top and bottom lines.
• "In the past five to seven years, their mandate has significantly changed,” observes Mistry. “They've had to become more of a tactician, economically, on both sides of the equation.”
• This is being reinforced by the move from pilots into core infrastructure, with workflows becoming the coin of the realm: “Companies are going straight into a validated strategy, whereby their tech stack needs to show immediate efficiency improvements, anything which is attributable to something economically and which hits the bottom line,” he notes.
• Indeed, leading companies are emphasizing production at-scale. Eli Lilly’s Diogo Rau, EVP and chief information and digital officer, said at a recent conference that the drugmaker’s unified AI platform, Cortex, “grew from offering one to now 52 models, includes ~60 enterprise-grade applications, and is used by almost every non-manufacturing employee across the globe.”
• That leads to big shift No. 2, the mandate for every employee to innovate with AI, including helping to build AI’s semantic data layer by infusing it with their own domain expertise. At the same conference, Johnson & Johnson said it built a governed semantic layer on AWS — “using knowledge graphs, LLM-powered metadata enrichment, and business glossaries to give AI models the domain vocabulary and entity relationships needed to interpret commercial data correctly.”

5. What’s next

Now that it’s co-locating with AI engineers, how can pharma work best with these companies to realize real-world value from agentic AI, especially in commercial areas where it's starting to make inroads (think targeting, sales-force incentives, and analysis of Rx abandonment)? Mistry says the following strategic enablers will be pivotal over the next two to three years:

• Reimagining workflows: As operational leaders strive to rethink work in their domains, they need to stay focused on, “What is something completely different from what they were doing before that's AI net-new?”
• Context engineering: Mistry compares building pharma’s AI stack to installing wiring and piping for a new house. AI depends on the right data — including a retrieval layer, or RAG — that it can reason over, not just query. The semantic layer is the pipeline needed to turn a general-purpose LLM into a pharma commercial ops expert.
• Organizing around human-AI process: For agentic AI to become the “operating layer” for commercial workflows, managers will need to assess what repeatable processes can shift from human-coordinated to AI-orchestrated — and overcome cultural and organizational barriers to adoption.
• Investing in agnostic systems that continuously learn: Mistry says organizations need to ensure that what they’re investing in won’t become obsolete in the next six or eight months and will “evolve through the landscape and, hopefully to some degree, be agnostic to different models...Are you pushing all of the different vendors and solutions to show you that?”