challenges to r&d efficiency
Make it real. Smart bet, only option, or both? Biopharma R&D turns to AI.
Table of contents 2 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
3 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Executive summary The pharmaceutical and biotechnology (biopharma) industry faces persistent, long- standing challenges to its R&D efficiency. Top among them are the high cost of drug development and elevated clinical failure rates. The return on R&D investments has undergone a steady decline to a now flat, but unacceptable level, making it increasingly difficult to bring successful new drugs to market. AI is now emerging as a powerful catalyst for breakthroughs in R&D, precisely in the areas where the industry has struggled most in recent years. By streamlining discovery, optimizing trial design, and enabling predictive insights, AI is poised to bring to biopharma R&D more agile, data-driven, and outcome-oriented processes. Converging technology including advances in biology, physics, and computational power are unlocking this breakthrough now. Our global research reveals that biopharma organizations recognize this potential. It finds that 82% of executives believe AI will fundamentally transform biopharma R&D, and 60% agree that companies failing to scale AI will fall behind in innovation and market relevance. Executives also see AI as a key enabler in early-stage R&D, where the most complex challenges reside, supporting the design of experimental approaches and refining models and prototypes. In fact, 63% anticipate that most new molecular entities (NMEs) will originate from AI-driven platforms within the next decade. Recognizing the transformative potential of AI, most biopharma organizations (79%) are actively developing strategies to integrate AI across their R&D value chain. For example: • In drug discovery, 74% of executives believe generative AI (Gen AI) holds significant potential. Our research reveals that target identification is the most widely adopted AI use case in the drug discovery phase, with 43% of organizations implementing it. Of these, 32% report productivity improvements with an average time savings of 28% compared to before adopting the technology. • In clinical trials, organizations are already using AI for predicting endpoints and dosing 4 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Executive summary regimens, selecting trial sites, and predicting adverse events and treatment responses. More than 60% affirm that Gen AI can substantially improve the efficiency and outcomes of clinical trials. • In regulatory submissions, AI accelerates processes by automating data compilation from internal and external sources, enhancing quality through predictive modeling, and embedding responses to anticipated regulator queries upfront. A strong majority (73%) agree that Gen AI has the potential to fundamentally transform regulatory submission and approval workflows. Among organizations using AI to prepare and submit regulatory documents, 37% report productivity improvements, with an average time savings of 19%. Organizations are also actively exploring and piloting AI agents in R&D, with nearly three in 10 already piloting initial use cases. Plans for growing levels of investment indicate that companies expect transformation to accelerate. It remains unclear, though, whether the industry will have the capacity to use AI for R&D in the most effective ways. Despite having established foundational data capabilities, many biopharma organizations remain underprepared in data readiness. Data quality, external data procurement and integration, and data sharing, among other areas, are top challenges. Operational and cultural readiness to scale AI also remains a significant hurdle. To unlock the full potential of AI in biopharma R&D, organizations must address these gaps. We conclude the report with recommendations for accelerating the adoption and scaling of AI: • Ensure top-down support with senior leadership buy-in • Assess organizational risk profile and define clear goals • Balance building core AI capabilities with strategic partnerships • Build a data- and digital-savvy workforce • Advocate for organizational data readiness and industry data standards. 5 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Who should read this report and why? This report is intended for C-suite executives at global pharmaceutical and biotechnology (biopharma) organizations. It offers recommendations to help senior biopharma leaders benchmark their AI maturity and understand the benefits that AI can bring to the drug discovery and development process. Given the potential that the technology has to transform the R&D value chain, this report will also be of high interest to executives across clinical and technological roles and functions. This report offers insights into how large and mid-sized biopharma organizations can implement AI and scale their AI use cases. This report is based on original findings from an industry survey of 500 senior executives (director level and above) at leading biopharma organizations across nine countries. All have revenue above $500 million and over half (61%) above $1 billion. Half of executives surveyed work in early-stage R&D (e.g., discovery, preclinical research) and half within late-stage R&D (e.g., clinical development, regulatory submission). We also convened an industry advisory board of two senior executives for this report. Advisory board members served as a sounding board from research development through analysis and report publication. In addition, we interviewed more than 10 senior biopharma executives at leading organizations. Please see the research methodology at the end of the report for more details. 6 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Key definitions Thank you to our advisory board members Our advisory board members provided highly valuable input on multiple occasions that helped us shape and develop our report. Anna Berg Asberg Senior Vice President and Global Chief Information Officer, Gilead Sciences Matthias Essenpreis Chief Technology Officer, Roche Diagnostics 7 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Key definitions In this research, when we refer to artificial intelligence (AI), we mean the full range of technologies that enable machines to learn, reason, and make decisions – either by analyzing data, generating content, or taking autonomous actions. We use AI as a collective term to reflect the full spectrum of capabilities. This includes, but is not limited to: • Machine learning (ML) : Systems that learn from data to make predictions or identify patterns. • Generative AI (Gen AI): A subset of AI that harnesses the power of extremely large models and massive scaling of data and computing power to plan, reason, and create generative features (i.e., new content based on patterns and data they have learned from) including text, image, video, and code. • AI agent: Any AI system that can interact with its environment, collect data, and use this to autonomously perform tasks to meet predetermined goals. • Agentic AI: The deployment of AI agents in a real-world environment where agents can detect signals, plan and reason, make autonomous decisions, and achieve set goals independently without human intervention – such as adjusting manufacturing processes in real time or planning and executing lab experiments. 8 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Discovery Preclinical testing Early-stage Early clinical trials (Phase 1 & 2) Phase 3 clinical trials Clinical manufacturing Late-stage Regulatory approval • Lab testing • Animal trials • Tech transfer • Validation • Phase 1 • Phase 2 • Dossier submission Chemistry, manufacturing, and controls (CMC) Source: Capgemini Research Institute analysis. Note: CMC is the process that ensures the drug can be reliably and consistently manufactured into a safe and effective medication that meets all required quality and safety standard; tech transfer is transferring knowledge, processes, and technology from R&D phase to clinical manufacturing stage to ensure the drug can be manufactured consistently, reliably, and in compliance with regulatory standards; validation ensures that the processes, procedures, and activities performed during the manufacturing of drug generate reproducible results that meet quality standards. The biopharma research and development (R&D) value chain as defined in our research covers drug discovery through regulatory approval/dossier submission. We acknowledge that the nomenclature within the value chain might differ by organization; however, for the purposes of this research, below is our classification to allow for standardization of responses in our survey. 9 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Quantum safety is on the radar of most organizations 01 AI is the biopharma industry’s prescription for ailing innovation 01 10 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
The growing imperative for AI-driven transformation Biopharma R&D productivity, or the average investment required per successful new drug, has been worsening since the 1950s. Eroom's Law observes that, from 1950 to 2010, the number of new drugs approved for a given, inflation-adjusted R&D spend fell by about 50% every nine years. Put another way, even after discounting for inflation, a new product cost biopharma 100 times more to develop in 2010 than six decades earlier. 1 More recent estimates from 2024 place the average cost of bringing a new drug to market at $2.6 billion, and taking 10 to 15 years on average. 2 Helping drive this daunting level of outlay is the roughly 90% failure rate in clinical trials. 3 The most common culprits are products that do not work well enough or that have unmanageably severe side effects. 4 Already high, and rapidly increasing, R&D inefficiency is unsustainable in normal times. “It is impossible to overestimate how vital it will be to see technological disruption revive biopharmaceutical R&D. Without it, the industry will find it increasingly difficult to fulfill its social mission, let alone attract investors interested in fair returns,” says Thorsten Rall, Executive Vice President, Global Life Sciences Industry Lead, Capgemini. The growing need for breakthrough and novel therapies makes the issue more pressing. Help, though, appears to be at hand. Several converging trends will allow AI to transform biopharma R&D. These include the simultaneous growing data breadth, advancing computing power, better algorithms, and increased industry readiness. In particular: • Expanding volume and variety of available biological and chemical data – The cost of sequencing has drastically declined, and omics data (genomics, proteomics, metabolomics) are now being linked at scale in public databases, such as the UK Biobank, National Center for Biotechnology Information, and Cancer Genome Atlas. • Improvements in computational infrastructure and advances in large language models (LLMs) – Cloud computing and high-performance “cheap” computing power, makes it feasible to train and deploy AI models on vast datasets. Such tools can now, among other activities, predict protein structures, generate novel drug-like molecules, optimize synthesis routes, and analyze unstructured R&D documents to better understand biology and chemistry at a systems level. Beyond LLMs, AI techniques like autoencoders enable efficient processing of complex biological and chemical data, complementing generative models to accelerate discovery. • Integration of AI across the R&D value chain – Companies are using these IT improvements to embed AI in every part of R&D, from target identification and molecule design, through preclinical safety and efficacy prediction, to clinical trial design. • Regulatory momentum – The FDA and EMA are actively engaging with AI models and tools. 5 6 • Industry and AI-native partnerships – Large pharmaceutical companies are partnering with AI-native start-up and biotech firms. Examples include those between Sanofi and Exscientia (now Recursion) and between Pfizer and Insilico Medicine. 7 8 In short, AI-driven R&D could provide the increasingly necessary transformation for biopharma R&D to continue to bring to market the new products society needs at prices it can afford. 11 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Companies will use AI to rebuild drug discovery and development, or quickly fall behind The competitive clock is ticking for deploying AI across R&D. Our expert interviewees, and a large majority of our executives surveyed, agree that widespread change has already begun and will quickly reshape this field in various ways. AI adoption within R&D has already started. Jim Weatherall, Chief Data Scientist, Biopharmaceuticals R&D at AstraZeneca, writes that, “Data science and AI are transforming R&D, helping us turn science into medicine more quickly and with a higher probability of success. We are applying AI throughout the discovery and development process, from target identification to clinical trials, to uncover new insights to guide our drug discovery and development.” 9 He is not alone in his thinking. Over eight in 10 executives believe that AI will fundamentally transform biopharma R&D. The expected widespread adoption of AI in Thorsten Rall, Executive Vice President, Global Life Sciences Industry Lead, Capgemini “It is impossible to overestimate how vital it will be to see technological disruption revive biopharmaceutical R&D. Without it, the industry will find it increasingly difficult to fulfill its social mission, let alone attract investors interested in fair returns” 12 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 1. The majority of executives agree that AI will transform biopharma R&D R&D springs from fear as much as from hope: 63% warn that organizations failing to scale AI will fall behind in innovation and market relevance (see Figure 1). As Brian Eden, Vice President, Global Life Sciences, Capgemini puts it, “AI-driven R&D transformation is more than an attractive choice. It is a basic strategic necessity. Companies that wait for others to figure out the new way of doing things will disappear.” A senior executive at a global pharmaceutical company believes, “In the next five years, companies will see internal process improvements and better outcomes from AI. Drug development will benefit from enhanced data generation and analysis, such as digital pathology (i.e., managing and analyzing information generated from digitized specimen slides) over a longer period.” Percentage of executives who agree with the statements AI will fundamentally transform biopharma R&D within the next five years Biopharma organizations that fail to scale AI will fall behind in innovation and market relevance 82% 63% 13 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Alexandre Malouvier Senior Director Scientific Affairs and Digital Innovation, MAPI Research Trust/AI Centre of Excellence at ICON plc “AI can revolutionize drug discovery. With better data and smarter selection, especially in small molecules and peptides, we can dramatically improve how we identify promising candidates." The forecast gains outweigh the perceived risks Such expectations of widespread AI-driven R&D transformation reflect a growing consensus on risk and reward. Of respondent executives, 63% agree that, within 10 years, AI-driven platforms will find most new molecular entities (NMEs) (see Figure 2). At the companies of those surveyed, on average today 12% of pipelines are AI-discovered targets. This share is expected to reach 34% in five years and 60% in 10 (see Figure 3). More generally, 61% say that the technology will dramatically accelerate drug discovery process. Bob Bradway, CEO at Amgen puts it simply: “Machine learning and AI better enable us to choose what we think can be the winning molecules.” 10 The industry impact looks set to be huge. Already, Sanofi’s new discovery engines, powered by AI, machine learning, and experimental validation, have unlocked seven novel drug targets in just one year. 11 Meanwhile, only 22% of executives agree that the cost and risk of implementing AI outweigh its current benefits for their R&D pipeline. These advances may take patience to demonstrate, but they will occur, says a data, digital, and 14 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
IT executive at a global pharmaceutical company: while efficiency gains observed to date “are more modest than initially expected. I believe benefits will improve as AI scales. Sure, there was initial hype around AI in R&D, but perceptions are now stabilizing as organizations focus on practical use case augmentation and overcoming challenges to realize real value.” A digital and data science executive at a US-based biotech company, also thinks that “AI will reshape biopharma R&D over the next five to 10 years, but not without challenges. As digital health tools become more widely adopted, they will provide high-quality, structured data, ideal for training machine learning models. [Then] we will see real productivity gains.” Figure 2. Most executives agree that AI will accelerate drug discovery Percentage of executives who agree with the statements The majority of new molecular entities (NMEs) will originate from AI-driven platforms within the next decade Expanding our use of AI will dramatically speed up the drug discovery process The cost/risk of implementing AI outweighs its current benefits in our R&D pipeline 63% 61% 22% Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. 61 % say that the technology will dramatically accelerate drug discovery process 15 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
AI will evolve beyond an efficiency tool to help steer the strategic future of biopharma R&D There was a consensus among survey respondents and interviewees that AI has the potential to impact each stage of the R&D lifecycle. For example, the data, digital, and IT executive at a global pharmaceutical company says that the technology will “revolutionize the biopharma space by accelerating early-stage drug discovery, unlocking hidden patterns in biomedical data, and enabling faster, more confident decision-making. Over the next five to seven years, we will [also] see shorter timelines from lab to bedside, smarter knowledge management, and more targeted upskilling, ultimately transforming healthcare R&D and making investments in this space more lucrative.” Alexandre Malouvier, Senior Director Scientific Affairs and Digital Innovation, MAPI Research Trust/AI Centre of Excellence at ICON plc, adds that, “AI can revolutionize drug discovery. With better data and smarter selection, especially in small molecules and peptides, we can dramatically improve how we identify promising candidates. That’s where my hope lies for the next five to 10 years.” Figure 4 shows our view on what the future of drug discovery could look like when reshaped by AI. Figure 3. The share of AI-generated targets is projected to reach 60% in the next decade Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Average share of pipelines that are AI-discovered targets, today and projected in five and 10 years Today 2030 2035 12% 34% 60% % of pipelines with AI-discovered targets 16 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
17 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Figure 4. AI will transform every element of biopharma R&D in the years ahead AI will transform the entire biopharma R&D value chain. Discovery Preclinical testing Clinical trials Clinical manufacturing Regulatory approvals • Comprehensive AI models with improved multimodal data management from diverse datasets such as real-world patient data, genomic, proteomic, environmental, and clinical data. • Gen AI models to design novel compounds with high binding affinity and drug-like properties. • Simulation of millions of compounds against biological targets, dramatically expanding exploration and reducing trial-and-error. • Advanced models to predict compound interactions with the human body using in vitro and in vivo data, improving biological understanding and reducing development risks. • Predictive toxicology and ADME (absorption, distribution, metabolism, excretion) modeling using advanced in silico simulations. • Robotics driving high-throughput analysis and cell-based assays, accelerating experimentation and reducing manual errors. • Automated mechanistic insights in pharmacokinetic-pharmacodynamic (PKPD) modeling, enabling faster and more accurate dose-response predictions. • Smarter patient matching to trials using EHRs, genomics, and social data. • Real-time data analysis enabling trials to adjust protocols dynamically. • Continuous device-generated outcome tracking to track real-time patient progress and treatment effectiveness. • Improved acquisition, cleansing, and management of clinical and treatment data, enhancing trial quality. • Sharper performance analytics, helping optimize therapies and accelerate development. • Seamless, one-click tech transfer powered by AI-driven automation. • Intelligent bots automating routine lab tasks, freeing scientists for strategic work. • Dynamic batch sizing tailored to trial demand and patient profiles. • Algorithms to fine-tune bioreactor conditions, raw material usage, and batch consistency for higher yield and quality. • Continuous monitoring and automated control of manufacturing processes to ensure precision, compliance, and efficiency. • Automated regulatory document generation. • NLP-driven submissions to streamline regulatory submissions and ensure compliance with evolving standards. • Post-market intelligence to support approvals and label expansions. • Real-time tracking and updating of regulatory information. Continued on next page 18 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Figure 4. AI will transform every element of biopharma R&D in the years ahead – Part 2 Continued on next page • Advanced discovery methods enabling quicker and more targeted treatments. • Personalized therapies based on genetic profiles. • Smarter drug repurposing/repositioning • Reduced reliance on physical assays. • Accelerated viable drug candidate discovery. • Smarter candidate selection, ensuring only the most promising therapies enter clinical trials, boosting pipeline efficiency. • In silico models and predictive simulations to minimize the need for animal studies and lab-based assays, saving time and cost. • Shorter trial timelines leading to faster patient recovery and quicker market access. • Smarter candidate selection and adaptive protocols to reduce dropout and improve trial outcomes. • Lower trial cost. • Consistent product quality, reducing variability and enhancing compliance. • Faster interpretation of trial outcomes, allowing timely adjustments and improved decision-making. • Faster submissions with consistency and no/less errors. • Improved transparency and responsiveness. • Regulatory guidance improving on the credible frameworks. AI will deliver significant value across each phase. Discovery Preclinical testing Clinical trials Clinical manufacturing Regulatory approvals 19 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute analysis. Organization-wide value proposition Human oversight will still be essential for governance and ethics. Increased in-house capabilities: less reliance on external partners. Smarter decisions: data-driven, high-quality choices. Lower failure rates: reduced R&D budget impact. Expanded access: more treatments for previously untreatable conditions. Empowered scientists: routine tasks automated; focus on complex oversight. CMO transformation: AI-driven business model evolution. Precision medicine: right drug, right target, right condition. Process optimization: improved internal workflows and outcomes. Faster delivery: quicker access to medicines. Unified teams: full lifecycle ownership in development and manufacturing. Figure 4. AI will transform every element of biopharma R&D in the years ahead – Part 3 Discovery Preclinical testing Clinical trials Clinical manufacturing Regulatory approvals 20 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
21 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Laying the groundwork for big bets on AI in R&D 02 22 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
Most biopharma companies are setting or beginning to implement strategy A majority of executives work for organizations that are formulating an organization-wide strategy for AI in R&D or are looking to roll one out (71%). Only 8% are actively implementing a strategy and roadmap (see Figure 5). Maturity here increases with company size. For example, just 2% of companies with $1–5 billion in revenue have begun implementation. For those with more than $20 billion in revenue, that figure reaches 41% – much higher, but still a minority. Justin Melnick, Director, Global Life Sciences Incubator, Capgemini says, “For years, large companies have been building strong teams that set them up to move quicker into the strategy phase. However, they face challenges on scaling that the smaller more agile firms do not. They need to commit to and drive rapid transformation at scale if they are to stay ahead of the more agile firms.” Our research assessed how advanced companies were on AI adoption in R&D. It found most to be at early stages, but ready to ramp up quickly. Justin Melnick Director, Global Life Sciences Incubator, Capgemini “For years, large companies have been building strong teams that set them up to move quicker into the strategy phase. However, they face challenges on scaling that the smaller more agile firms do not. They need to commit to and drive rapid transformation at scale if they are to stay ahead of the more agile firms.” 23 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 5. 71% of organizations are currently formulating a strategy for AI in R&D or preparing for rollout Percentage of executives who say their organization has a strategy for implementing AI in R&D Stage 0: We do not have a strategy for AI in R&D nor have any AI initiatives ongoing Stage 1: No organization-wide strategy, but some isolated AI initiatives exist in R&D Stage 2: We are currently formulating an organization-wide strategy for AI in R&D Stage 3: We have finalized our AI in R&D strategy and are preparing an implementation roadmap Stage 4: Our strategy and roadmap for AI in R&D are in place, and are actively being executed, with ongoing efforts to refine our approach in light of AI advancements 4% 17% 43% 28% 8% 24 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
A blend of internal development and external innovation is the most common way to build capabilities Slightly over half (52%) of executives report using a hybrid mix of in-house activities, partnerships, and acquisitions to develop AI capacity in R&D. Another 29% are primarily partnering with external AI-native companies. Only one in 10 organizations are relying wholly on internal capability creation (see Figure 6). Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 6. Half of organizations build AI in R&D capabilities internally and with external partners Percentage of organizations by primary strategy for building AI capabilities in R&D A hybrid approach (mix of in-house, partnerships, and acquisitions) Partnering with external AI-native companies (via licensing agreements) Building capabilities entirely in-house 52% 29% 10% 52 % of executives report using a hybrid mix of in-house activities, partnerships, and acquisitions to develop AI capacity in R&D 25 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Novartis is a good example of what hybrid development looks like. It has a dual-track AI strategy, building strong in-house platforms, talent, and tools while partnering with firms like Isomorphic Labs (an AI-first drug design company), 12 Generate:Biomedicines (an AI-driven therapeutic development firm) 13 and Microsoft 14 to tackle complex R&D challenges and develop core AI infrastructure. Novartis also co-leads MELLODDY (Machine Learning Ledger Orchestration for Drug Discovery), an AI platform shared among several pharma companies. AstraZeneca is also taking a hybrid approach. Besides building tools in house, it has: developed a protocol assistant with medical writers; 15 signed a $555 million deal with Algen Biotechnologies to use the latter’s AI platform for immunology research; 16 and partnered with BenevolentAI, an AI-native firm, for the discovery and development of new treatments for chronic kidney disease (CKD) and idiopathic pulmonary fibrosis (IPF). 17 AstraZeneca also collaborates with Stanford Medicine to tap into its research and tech expertise. 18 Roche, a Swiss pharma multinational, adopts a similar model. Its subsidiary Genentech’s research and early development group (gRED) operates independently within Roche, preserving a startup-like culture while collaborating on new medicines. 19 Roche also partners with experts in drug discovery, including a four-year AI-driven collaboration with Recursion. This partnership recently produced a second neuroscience-focused whole-genome map, offering a novel approach to identify targets and pathways tackling key challenges in neuroscience drug discovery. 20 26 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Biopharma organizations expect that today’s investments in AI in R&D are just the beginning Over two-thirds of executives (68%) expect their corporate AI investment in R&D to increase between 2025 and 2026. Of all companies surveyed, the average forecast increase is 12%. Larger companies are making bigger proportional bets on the technology. For example, in 2025, surveyed organizations with annual revenues between $500 million and $1 billion dedicated 34% of total AI investment to R&D, for an average total of $1.8 million that year. For companies with more than $20 billion in annual revenue, the equivalent figures are 47% and $70 million (see Figure 7). This disparity looks set to grow. Over eight in 10 of the larger companies (81%) expect to increase investment in AI, well above the 60% average among the smallest companies ($500 million to $1 billion in revenue), likely because larger organizations can spread platform investments across a broader portfolio, making size a significant competitive advantage. Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=239 pharmaceutical and biotechnology organizations. Figure 7. The share of total AI investment dedicated to R&D increases with company size Average share of total AI investment dedicated to R&D, 2025 $500 mn–1 bn $1–5 bn $5–20 bn $20 bn+ Company size (annual revenue USD) 34% 36% 50% 47% 27 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
AI investment is most common in drug discovery but should be the norm across the value chain in five years Of surveyed organizations, 57% currently invest in AI for drug discovery, while 41% plan to do so in the next three to five years. This is already having marked effects. The number of biologics and vaccines developed at Sanofi with the aid of AI has nearly doubled since 2019. 21 Looking ahead, Novartis will shift drug discovery from AI-enabled to fully AI-driven across R&D. 22 Meanwhile, AstraZeneca’s Centre for Genomics Research aims to analyze up to two million genomes by 2026, using data science and AI to accelerate insights. 23 While fewer companies are investing in AI elsewhere in the R&D value chain, the overwhelming majority plan to do so within five years. By then, 11% or fewer will not have deployed AI in any given R&D phase (see Figure 8). 28 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=450 pharmaceutical and biotechnology executives. CMC: Chemistry, manufacturing, and controls. Figure 8. The vast majority of organizations are currently investing, or planning to invest, in AI across all value chains stages Percentage of organizations investing in AI by phase of the R&D value chain Drug discovery Preclinical testing Clinical trials CMC/Clinical manufacturing Regulatory submission/approval Currently investing Planning to invest in the next 3–5 years No plans to invest/not sure 57% 41% 2% 37% 54% 9% 46% 44% 11% 50% 42% 8% 29% 63% 8% 29 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
How AI is reshaping parts of the R&D value chain today 03 30 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
The impact of different forms of AI will vary across R&D phases Helping to drive more widespread investment is the impact of AI already apparent from use cases in every R&D phase.This section takes a closer look. AI’s utility within the R&D value chain varies depending on the specific application. Most surveyed executives, for example, see Gen AI transforming drug discovery (74%) and regulatory processes (73%). For pre-clinical testing, only 39% say Gen AI will have that impact. Instead, natural language processing (NLP) is expected to transform that field (51%). Machine learning, meanwhile, is widely expected to have an important effect on drug discovery (72% of respondents) and clinical trials (66%). Overall, more than half say that at least one type of AI, and usually more, will be transformative (see Figure 9). What this will look like in practice will differ by R&D phase. Flavius Martin, M.D., Executive Vice President, Research, Gilead Sciences says, “The use of generative AI in drug development, enabled by people, science, and other new technology, has shown potential to accelerate the discovery of molecules for challenging targets.” 24 Meanwhile, AI can accelerate regulatory submissions by automating data compilation across internal and external sources. It can also enhance quality by anticipating regulator queries and embedding responses upfront. As Boris Braylyan, Vice President and Head of Information Management at Pfizer explains, “In the future … [w]e may … be able to improve our submissions by predicting in advance what regulators are likely to ask, and coming prepared with those answers.” 25 Novartis is already using AI to ease the manual load of regulatory compliance and reporting. It has invested in Yseop to use that firm’s NLP technology in order to automate clinical report writing. 26 74 % of executives see Gen AI transforming drug discovery 31 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=478 pharmaceutical and biotechnology executives. *CMC: Chemistry, manufacturing, and controls. Figure 9. Seven in 10 executives believe Gen AI will improve drug discovery and regulatory submission Percentage of executives believing different forms of AI have transformation potential to improve R&D productivity in the R&D value chain Drug discovery Preclinical testing Clinical trials CMC/Clinical manufacturing Regulatory submission/approval ML NLP 72% 74% 45% 47% 39% 33% 49% 51% 66% 60% 62% 51% 32% 41% 52% 36% 24% 62% 73% 52% Gen AI AI agents 32 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Most organizations are exploring or piloting AI agents in R&D Of respondents’ organizations, 38% are exploring potential applications of AI agents in R&D. An additional three in 10 are piloting initial use cases. Only a very small proportion (a mere 6%) say that they have no plans for AI agents in this (see Figure 10). The reason for this interest is simple. A data, digital, and IT executive at a global pharmaceutical company explains that “AI agents have the potential for higher efficiency and user- friendly outcomes.” Diverse challenges, though, make finding value a work in progress. An executive from a healthcare manufacturer says, “We are currently deploying internally developed agents in a secure cloud environment … and not using store-bought AI agents due to security and change management concerns.” A digital and data science executive at a US-based biotech company adds that such agents are “definitely on our radar, [but] it is early days, and we are still evaluating where agentic AI could add real value beyond what chatbots offer. One idea … is automating the generation of SDTM (Study Data Tabulation Model) datasets from raw clinical data.” Sheetal Chawla, Executive Vice President and Head Life Sciences North America, Capgemini, advocates that governing AI agents is mission-critical: “As AI agents evolve beyond pilots into real- world application, capable of sensing, learning, and acting across complex R&D workflows, the conversation must shift from feasibility to responsibility. Governance, transparency, and trust are essential. Leading organizations will embed AI agents into their scientific core while maintaining rigorous oversight and a clear sense of purpose.” 33 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Sheetal Chawla Executive Vice President and Head Life Sciences North America “As AI agents evolve beyond pilots into real-world application, capable of sensing, learning, and acting across complex R&D workflows, the conversation must shift from feasibility to responsibility. Governance, transparency, and trust are essential. Leading organizations will embed AI agents into their scientific core while maintaining rigorous oversight and a clear sense of purpose.” 34 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 10. Approximately a third of organizations (34%) are piloting or already using AI agents in R&D Percentage of organizations by stage of AI agent implementation No plans to explore AI agents in R&D Planning to adopt AI agents in R&D within the next 2–3 years Planning to adopt AI agents in R&D within the next year Currently exploring potential applications of AI agents in R&D Piloting initial use cases of AI agents in R&D Already using AI agents in R&D for specific tasks or decision support 6% 11% 12% 38% 29% 5% only 6 % of organizations do not have plans to explore the use of AI agents in R&D 35 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Those organizations actively adopting AI are seeing productivity improvements across R&D As noted earlier, organizations have begun implementing AI in every phase of the R&D value chain, a trend set to continue. This activity involves exploration of diverse use cases to unlock new efficiencies and accelerate innovation. A senior executive from a biopharmaceutical company says, “I anticipate that AI will substantially decrease the time required from target identification to patient delivery. Additionally, AI is expected to automate manual tasks, facilitate increased in-house operations while reducing reliance on external resources, lower error rates, and enable earlier and more accurate decision- making. This will all result in improved patient outcomes.” Specific examples abound. AstraZeneca is deploying AI and data science to boost R&D productivity significantly, including supporting over 240 global trials and launching multiple pilots aimed at streamlining processes. 27 Similarly, Novartis has implemented AI across more than 100 use cases company-wide, driving innovation in R&D to accelerate the development of novel therapies and drugs. 28 Meanwhile, Sanofi’s large language model, CodonBERT, trained on 10 million mRNA sequences, helped the company to reduced mRNA design time by 50%. This savings in turn accelerates development of mRNA vaccines and therapies. 29 Indeed, many organizations are using AI in ways that help with different R&D phases simultaneously. For example, Gilead is deploying Gen AI to accelerate drug discovery for complex targets, which it expects will also cut months off the regulatory target assessment timeline. 30 In the following sections, we detail a few of the most prominent use cases implemented today at each step of the R&D value chain. Drug discovery In this field, 43% of surveyed organizations use AI in target identification (finding the specific molecular target of a disease or condition), making it the technology’s most common drug discovery use. A further 23% are piloting its application for this purpose (see Figure 11). To give just one example, Sanofi is using AI to identify key targets for both HIV and cancer. It is then applying the technology to create single, complex molecules with multiple antibodies which simultaneously act on three of these targets in the pathways of one or more diseases. 31 36 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 11. Over four in 10 organizations have implemented AI for target identification in drug discovery Percentage of organizations by AI use case implementation stage in drug discovery Target identification Biomarker identification Side effect identification Compound screening Predicting drug-likeness De novo drug design Drug interaction simulations Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 3% 30% 23% 43% 6% 22% 41% 31% 4% 42% 24% 29% 7% 35% 31% 25% 4% 53% 21% 20% 10% 38% 31% 19% 9% 58% 14% 18% Our survey reveals that important benefits from AI’s use in drug discovery are already becoming apparent. AI drives real value Target identification 28% time savings on average Biomarker identification 22% time savings on average Compound screening 20% time savings on average 37 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
• Of the 214 organizations that have implemented AI for target identification, 32% report a resultant productivity growth. The time needed for the process, compared to before adoption, has dropped 28%, or more than three months per year. For example, if it took 12 months pre-AI to identify targets, after using AI, the timeline shortened to 8.6 months. Merck, a Germany-based science and technology company, has gone dramatically further. It reduced chemical identification time from six months to just six hours using its AI-powered R&D assistant. 32 • Of the 123 organizations that have implemented AI for compound screening, half see better productivity improvement. The average time saved is 20% of that needed before AI. For example, if it took six months pre-AI to conduct compound screening, after using AI, the timeline shortened to 4.8 months, saving 2.4 months. Bristol Myers Squibb’s (BMS) AI-enabled predict- first strategy shows what this looks like in practice. BMS now forecasts success for most small molecules before synthesis, up from just 5% in 2021. The results are significantly improved speed and quality across their programs. 33 • Of the 154 organizations that have implemented AI for biomarker identification, 25% have already realized better productivity improvement. Here, the average percentage time saved is 22%. For example, if it took 24 months pre-AI to identify biomarkers, after using AI, the timeline shortened to 18.7 months, saving over five months. Preclinical testing Executives report slower rollout of AI across preclinical testing than in drug discovery. Nevertheless, a third have already implemented it for image annotation and segmentation. The near future will see more activity, with large numbers engaged in piloting AI use cases across preclinical testing, including four in 10 organizations doing so for bioactivity modeling (i.e., predicting how compounds interact with biological targets) (see Figure 12). 38 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 12. A third organizations have implemented AI to assist with image annotation and segmentation within preclinical testing Percentage of organizations by AI use case implementation stage in preclinical testing Image annotation and segmentation: AI-assisted labeling and segmentation of in vitro and in vivo imaging data to support toxicity, efficacy, and biomarker studies ADME analysis: Predict and optimize absorption, distribution, metabolism and excretion properties for pharmacokinetics Animal model selection: AI analyzes historical data, omics, and literature to recommend the most predictive animal models for specific disease pathways or compound classes Intelligent document management systems Bioactivity modeling: Predicts how compounds interact with biological targets to prioritize candidates Toxicity prediction: ML models forecast potential toxicity of compounds, reducing reliance on animal testing Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 3% 29% 34% 33% 6% 42% 24% 27% 7% 35% 30% 26% 3% 57% 17% 22% 6% 35% 41% 18% 8% 50% 25% 15% 39 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Clinical trials A little over a quarter of executives report that their organizations are using AI to help automate clinical trial processes, detect adverse events, and predict treatment responses. Again, these figures reflect the first stages of wider, industry-wide adoption. For example, 40% are piloting AI to predict endpoints and dosing regimens, with 42% doing so for selecting trial sites (see Figure 13). BMS (Bristol Myers Squibb) shows the possible benefits. The company is streamlining clinical trials and enhancing patient outcomes by applying AI/ ML to generate disease insights, patient profiles, and predictive models. This information enables smarter trial design, endpoint optimization, and subgroup targeting. 34 According to the company’s former SVP of Global Biometrics and Data Sciences, replacing 25% of standard-of-care patients with AI modeling can cut recruitment costs and shorten trial timelines by 10–15%. 35 As with any new technology, how companies use AI in trials will depend on inventiveness. A digital and data science executive at a US-based biotech company says that “One promising use case is forecasting … whether clinical trial sites will meet recruitment targets using real-world and operational data. If a site falls short, we identify alternative centers to pick up the slack, accelerating recruitment timelines.” 40 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 13. Nearly three in 10 organizations have implemented AI to automate clinical trial processes and predict adverse events Percentage of organizations by AI use case implementation stage in clinical trials Automation of administrative processes Predictive analytics: Adverse event detection Predictive analytics: Treatment responses Imaging biomarker extraction: AI analyzes imaging data collected during trials to identify predictive or prognostic markers Patient recruitment: AI identifies eligible patients from EHRs and other data sources, improving trial efficiency Trial design optimization: Simulates trial outcomes to design more effective and adaptive trials Predictive analytics: Endpoints and dosing regimen identification Site selection: AI analyzes historical trial performance, patient demographics, investigator experience, and operational metrics to recommend optimal trial sites Real-world evidence analysis: ML analyzes post-market data to assess drug safety and effectiveness Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 6% 33% 33% 28% 8% 34% 30% 28% 8% 42% 23% 27% 10% 42% 22% 26% 3% 33% 37% 26% 9% 43% 22% 25% 6% 29% 40% 24% 2% 34% 42% 22% 5% 54% 20% 20% 41 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
to be cont. CMC/clinical manufacturing The most commonly implemented AI use case within CMC (chemistry, manufacturing, and controls) and/or clinical manufacturing is predictive maintenance (38%). Early-stage activity is also common in other areas: over a third are piloting AI for supply chain forecasting (37%) and for multivariate analysis for validation (36%) (see Figure 14). If these numbers seem low, that is only in comparison to AI adoption in other R&D phases. As the data in the figure show, over half of companies have implemented or are piloting use cases in five different areas of clinical manufacturing. 42 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
43 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 14 (Part 1). Nearly four in 10 organizations have implemented AI for predictive maintenance in clinical manufacturing Percentage of organizations by AI use case implementation stage in CMC/clinical manufacturing Predictive maintenance: ML predicts equipment failures to reduce downtime Supply chain forecasting: AI improves demand forecasting and inventory management Visual inspection of manufacturing processes: Computer vision models detect anomalies, ensure compliance, and support quality control through real-time image analysis AI-powered knowledge transfer and decision support: Synthesize SOPs, batch records, historical data and deviation for smoother tech transfer, training, troubleshooting across sites, while also providing intelligent decision support Adaptive control strategies: AI adjusts control limits dynamically based on historical performance and risk models Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 6% 30% 25% 38% 8% 26% 37% 28% 8% 34% 30% 28% 8% 35% 33% 23% 6% 48% 23% 23% Continued on next page 44 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Figure 14 (Part 2). Nearly four in 10 organizations have implemented AI for predictive maintenance in clinical manufacturing Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Percentage of organizations by AI use case implementation stage in CMC/clinical manufacturing Automated documentation analysis: NLP models extract and compare critical process parameters (CPPs) and quality attributes (CQAs) across sites and batches Multivariate analysis for validation: AI identifies hidden correlations between process parameters and product quality Digital twins for process replication: AI-driven simulations replicate manufacturing processes to ensure consistency across sites Process optimization: AI monitors and adjusts manufacturing parameters in real time for quality control Real-time validation monitoring: ML models flag anomalies during validation runs, reducing manual review Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 8% 53% 17% 21% 4% 38% 36% 21% 5% 46% 29% 19% 7% 47% 27% 18% 5% 56% 22% 16% 45 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
46 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 15. Three in 10 organizations have implemented AI to assist with translation and customization in regulatory submission and approval processes Percentage of organizations by AI use case implementation stage in regulatory submission/approval Multilingual translation and local customization Regulatory document preparation and submission: Automate document generation, data standardization, submission readiness checks Regulatory intelligence and strategy: Trend analysis in approvals, rejections, and labeling changes, competitor benchmarking Review and compliance support: Automated quality checks, audit trail analysis Post-submission monitoring: Labeling and safety updates, regulatory lifecycle management Not exploring this use case Assessing/developing strategy Testing/piloting Implemented 2% 37% 29% 31% 3% 49% 21% 26% 5% 32% 40% 22% 9% 52% 25% 14% 13% 55% 18% 14% Regulatory submission/approval So far, implementation of AI use cases in regulatory matters has focused largely on administrative processes. Of those surveyed, 31% say that their organizations already deploy the technology to support multilingual translation and local customization. Similarly, 26% use it to help with document preparation and submission. Looking ahead, though, a shift toward strategic support is underway. Forty percent of organizations are piloting the use of AI for regulatory intelligence and strategy (see Figure 15). 47 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
The value of even these administrative applications should not be underestimated. Of the 129 organizations that have implemented AI for preparing and submitting regulatory documents, 37% are already realizing productivity improvements. The average cut in the time needed for this process is 19%. For example, if it took 12 months pre-AI to complete regulatory documents, after using AI, the timeline shortened to 9.7 months, saving over two months. Eli Lilly, for example, with an external partner, implemented an automated process to draft patient narratives for clinical study reports using natural language generation. In one year, this saved 10,000 hours of writing and reviewing time, allowing teams to focus on more complex, scientific work. 36 Over a third of biopharma companies have seen productivity improvements from using AI to prepare for regulatory submission; 19% time savings on average. 48 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Q. What is your company's approach to AI? We are embedding AI across the company in a smart, purpose-driven way. AI is transforming how we innovate in R&D, optimize business processes and operations, and engage with patients. We have a two-way approach: we form strategic partnerships with leading AI companies and also build internal platforms, talent, and tools. This combination accelerates innovation, allows us to tailor solutions to our needs, and ensures that the expertise we gain is embedded within our company. Q. What AI use cases are being implemented in R&D ? We are leveraging AI throughout the R&D process. In the pre-clinical phase, AI helps us better understand disease biology and identify promising drug candidates; in the clinical phase, to help target populations and to design intervention studies; and in the development of digital therapeutics and devices to enable continuous monitoring. Insights from a senior biopharma leader at a global pharmaceutical company Source: Capgemini Research Institute interview, September 8, 2025. Q. How are you preparing your people for AI? Our people are at the heart of our AI transformation. We invest substantially in upskilling our workforce, ensuring all employees are equipped to use AI in their daily work. Initiatives like the rollout of M365 Copilot and dedicated training programs help build an AI-enabled, future-ready workforce. Human oversight remains essential; our experts are empowered to oversee and interpret AI outputs, ensuring responsible and effective use. Q. How are you preparing your people for AI? Our people are at the heart of our AI transformation. We invest substantially in upskilling our workforce, ensuring all employees are equipped to use AI in their daily work. Initiatives like the rollout of M365 Copilot and dedicated training programs help build an AI-enabled, future-ready workforce. Human oversight remains essential; our experts are empowered to oversee and interpret AI outputs, ensuring responsible and effective use. 49 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Barriers to biopharma realizing AI’s potential 04 50 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
While executives almost universally foresaw the application of AI across R&D at their organizations, our research highlights a disparity between such expectations and capacity to implement the technology. Data accessibility and integration are key weaknesses Our research asked about preparedness across 11 different data elements needed to support AI’s use in R&D. For eight of these, fewer than half of respondents believe that their organizations are adequately prepared (see Figure 16). Those where a majority say they are prepared are, not surprisingly, basic building blocks for AI: • Data foundations (i.e., having the infrastructure such as data lakes, vector databases, metadata tagging, compute power to support scalable AI applications in R&D) • Data governance (i.e., ensuring clear policies and oversight for data stewardship, lineage, compliance such as Good [x] Practice [GxP] and HIPAA, and ethical use in AI) • Data security and compliance (i.e., having robust controls to ensure data integrity, confidentiality, and traceability in line with Good Manufacturing Practice [GMP], GxP, and other regulatory standards for AI-driven R&D). Although most companies seem to have these pieces in place, it remains worrying that, even in such fundamental areas, more than four in 10 executives acknowledge that their organizations are not prepared. While data quality, volume, and variety are issues for over half of companies, the most widespread issue is data accessibility and integration, where only 22% describe their organization as currently prepared. An executive from a healthcare manufacturer explains that “a unified data approach is essential for scaling AI, but asset and data source diversity remain significant hurdles, especially for contract manufacturing organizations. Our company's data is often client-owned so we are primarily concerned with how AI is used internally for GMP (Good Manufacturing Practice) decision-making and data sharing.” Biopharma organizations understand the pressing need to address these weaknesses. Most expect to make extensive investments to rectify deficiencies over the next one to three years. Of surveyed executives, 73% name improved data accessibility and integration as the top priority to support AI-driven R&D. Next comes a strong focus on data security and compliance (65%). Eli Lilly, to cite one example, is making a significant investment in large-scale, high-quality data accessibility through Lilly TuneLab, an AI platform developed with global tech and AI/ML experts. TuneLab provides biotech firms with access to drug discovery models trained on over $1 billion in proprietary research data, helping unlock scientific insights, enable smarter early decisions, and improve success rates. 37 AstraZeneca, meanwhile, is putting money into stronger data foundations, using AI to drive insights across R&D and enhance IT infrastructure for tracking new solutions. The company is building in-house data products, partnering with external vendors, and has created a cross-value chain to identify key data assets and improvement areas. 38 51 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 16. Most organizations are underprepared on data accessibility and integration Percentage of executives who agree that their organization is prepared today Data foundations Data governance Data security and compliance Data exploration and analysis Data quality and completeness External data procurement and integration Data sharing Data variety Data volume Agentic orchestration Data accessibility and integration 58% 54% 53% 46% 44% 38% 35% 34% 30% 26% 22% 52 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
The industry is at a critical juncture; however, most biopharma organizations are failing to build comprehensive data capabilities To fully harness AI’s potential in R&D, biopharma organizations must be data-ready. Yet, our survey reveals that most organizations lag in critical dimensions of data readiness. Figure 17 highlights the key dimensions essential for unlocking AI’s potential in biopharma. 53 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute analysis. Figure 17. Critical dimensions essential for unlocking AI’s full potential in biopharma 1 2 3 4 5 6 7 8 9 10 11 Data volume Sufficient and relevant biological, chemical, clinical, and real-world data to train and fine-tune AI models Data quality and completeness Accurate data, comprehensive, and representative of diverse patient populations and experimental conditions Data sharing Secure mechanisms to source and share data with external partners (e.g., contract development and manufacturing organizations, contract research organizations, academic collaborators) Data exploration and analysis Tools and capabilities to perform exploratory data analysis to identify patterns, biases, and insights relevant to drug development Agentic orchestration Strategies to manage and coordinate multiple AI agents (e.g., for hypothesis generation, molecule design, trial simulation) effectively Data security and compliance Robust controls to ensure data integrity, confidentiality, and traceability in line with Good Manufacturing Practice (GMP), Good [x] Practice (GxP), and other regulatory standards for AI-driven R&D Data variety Datasets capturing the complexity of R&D, including multi-omics, imaging, EHRs, trial data, and unstructured scientific literature Data accessibility and integration Seamless access to structured and unstructured data across silos (e.g., discovery, preclinical, clinical) and can integrate it into AI workflows Data governance Clear policies and oversight for data stewardship, lineage, compliance (e.g., GxP, HIPAA), and ethical use in AI Data foundations Infrastructure (e.g., data lakes, vector databases, metadata tagging, compute power) to support scalable AI applications in R&D External data procurement and integration Mature processes to acquire, validate, and integrate external datasets (e.g., real-world evidence, omics data, literature) into our R&D pipelines, including procurement and governance of external models Data enablers Data enablers are structural and foundational capabilities that make data accessible, interoperable, and scalable for advanced analytics. Data behaviors Data behaviors are cultural norms and operational practices that define how data is governed, shared, and applied responsibly across the enterprise. 54 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Data conscious 21% Data novices 36% Data enabled 6% Data frontrunners 37% Data behaviors Data enablers We organized the critical data dimensions into two pillars: 1) data enablers and 2) data behaviors, and assessed the maturity of biopharma organizations in enabling AI in R&D. Our analysis reveals that 37% of organizations qualify as data frontrunners (see Figure 18), actively investing across both pillars and positioning themselves to lead with accelerated innovation cycles, enhanced productivity, and significant cost efficiencies. In contrast, 63% of organizations are lagging on one or both pillars, underscoring an urgent need to strengthen overall data readiness, including technology, tools, governance, skills, and cultural adoption. *The cutoffs for the four quadrants were established based on the average scores for the two axes: data enablers and data behaviors. Frontrunners are those who score above average on both axes. Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 18. The majority of biopharma organizations need to enhance data capabilities 63 % of organizations are lagging on one or both pillars of critical data dimensions 55 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute analysis. Organizations in this cohort are at an early stage with minimal maturity, and face significant limitations in both technical and cultural dimensions. Their enablers are weak, characterized by limited data availability, fragmented infrastructure, poor integration, and immature compliance processes. Behavioral aspects are equally underdeveloped, with minimal governance, restricted data sharing, and basic analytical capabilities. Consequently, these organizations are only engaged in exploratory efforts with few formalized practices, leaving them far from realizing the potential of AI-driven innovation. Data novices Key characteristics of cohorts Cohort Description Organizations in this cohort are strong in enablers but evolving in behaviors, and have established a modern technical foundation, including advanced infrastructure, cloud adoption, comprehensive data catalogs, robust security measures, and seamless integration. However, their behavioral aspects such as governance frameworks and data-driven culture are still developing, with limited orchestration and analytics practices in place. As a result, these organizations are technically prepared for AI workflows but require cultural maturity to fully capitalize on their capabilities. Organizations in this cohort have strong behaviors but limited enablers, exhibiting a solid foundation in governance, ethical compliance, and secure data sharing, with exploratory analytics tools beginning to emerge. However, their technical backbone such as infrastructure, system integration, and external data processes remains in the early stages of development. As a result, these organizations are laying the groundwork for AI adoption but lack the depth and scalability needed to fully realize its potential. Organizations in this cohort are leading across all dimensions, demonstrating both strong enablers and behaviors. On the enabler side, they possess rich, diverse, and high-quality data, seamless integration across systems, advanced infrastructure, and robust security and compliance frameworks, complemented by well-defined external data processes. Their behaviors reflect a culture of robust governance, secure data sharing, exploratory analytics, and effective orchestration of AI agents. Together, these capabilities enable scalable AI-driven R&D and foster continuous innovation, positioning such organizations at the forefront of biopharma transformation. Data enabled Data conscious Data frontrunners 56 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Few organizations are operationally ready to scale AI A significant majority of executives (76%) believe that AI will reshape the roles of scientists in R&D; however, under half (45%) agree that their organization is culturally and operationally ready to scale AI in R&D (see Figure 19). Bigger companies appear to be taking the lead: 62% of respondents at companies with $20 billion or more in revenue say that their organizations are ready to scale the technology. Getting culture right is crucial because AI requires people to make it work properly – a point sometimes lost amid talk of technological advances. Ron Weathermon, Head of Medical Affairs Strategy at Teva, explains, “There is an ongoing need for human oversight in drug development. While AI can accelerate data modeling and patient journey mapping, human input remains essential for understanding patient needs and validating models.” AstraZeneca’s change management approach has been to partner with different groups of employees and seek top-down support. The company holds “mini-SCOPEs,” knowledge sharing sessions to showcase different use cases and technologies. Learning modules and accreditation are also offered to employees at the gold, silver, and bronze levels to encourage them to continuously upskill on AI. 39 Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Figure 19. Fewer than half of organizations agree they are ready to scale AI in R&D Percentage of executives who agree with the statement AI will fundamentally reshape the role of scientists in biopharma R&D, shifting focus from experimentation to orchestration Our organization is culturally and operationally ready to scale AI in R&D 76% 45% 45 % of respondents agree that their organization is culturally and operationally ready to scale AI in R&D 57 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
How biopharma organizations can accelerate and scale AI in R&D 05 58 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
Based on the results of our primary research, key practices that distinguish data frontrunners, and our experience working on the ground in partnership with biopharma organizations, we share five key recommendations to help accelerate the adoption and scaling of AI. Broadly speaking, organizations must think beyond short-term use cases and consider how AI is enabling and reimagining the entire industry in the longer term. Ensure top-down support with senior leadership buy-in Many interviewees stressed the importance of a clear signal from senior leaders about the high priority they give to using AI to improve every aspect of R&D. In fact, nearly half (47%) of biopharma data frontrunners believe their organization will not be able to discover new therapies without the use of AI within the next five years, compared to 35% of data novices. To achieve meaningful results, leadership must provide specific guidance, realistic timelines, and sustained support to cross-functional teams working in this area. Senior executives should also foster a culture that embraces experimentation, collaboration, and continuous learning. Several interviewed experts see support from the top as essential for success in deploying and scaling AI. A senior executive from a biopharmaceutical company says, “Our AI strategy has advanced quickly thanks to support from the board and leadership. Strong top-down leadership and centralized decision-making are crucial for speeding up AI adoption.” A senior executive from a global pharmaceutical company, “Scaling AI typically starts with experimentation and requires top- down support for investment and process redesign.” Those in top corporate roles can drive AI adoption better if frontline employees keep them informed. Teams, by clearly communicating how AI can drive innovation, improve productivity, and enhance competitiveness, can shift leadership focus from immediate profitability to sustained value creation. Ways to do so include presentations on topics such as use cases, pilot outcomes, and industry benchmarks that demonstrate tangible impact. Encouraging open dialogue and sharing insights from cross-functional teams helps build trust and alignment. 1 59 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Assess organizational risk profile and define clear goals A comprehensive risk profile for AI adoption should move beyond technical risks and encompass ethical, regulatory, and strategic considerations throughout the entire R&D value chain. A risk position should be established based on the unique culture of the organization. That culture shapes the risk position by influencing a company's willingness to take risks, its ability to identify and respond to threats, and its overall approach to risk management. For example, a culture built on transparency and accountability fosters better decision-making, adaptability, and proactive risk management. A risk-averse culture, or even a high-risk 2 culture, can lead to overlooking risks or taking on too much, with all the possible attendant damage. Our research reveals that only 15% of biopharma data frontrunners believe the risk of implementing AI outweighs its current benefits to their R&D pipeline, compared to over a third (34%) of data novices. A potential explanation could be that biopharma data frontrunners perceive AI risks as manageable because they have the infrastructure, governance, and cultural readiness to mitigate them – while data novices see those same risks as barriers due to poor data foundations and limited organizational preparedness. This is further evidenced in our finding that 70% of data frontrunners have established ethical guidelines specifically tailored to the use of AI in R&D, compared to 49% of data novices. To realize the full potential of AI, organizations must develop a culture of innovation and cross-functional collaboration which, according to Ron Weathermon of Teva, “encourages experimentation and learning from failure.” Setting and defining clear, culturally aligned goals is also critical. For example, 60 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
perhaps a more risk-averse culture will focus on AI use cases considered “quick wins” such as summarizing study results, drafting protocols, or extracting key findings from research papers. A more risk-tolerant culture may focus on innovative, higher investment, higher-value use cases such as designing entirely new molecular entities. Ron Weathermon of Teva recommends that “Organizations start with small, validated AI interventions and scale gradually, not rapid, large- scale implementations without clear objectives and validation.” Executives should also appreciate that AI can help lower risk levels if deployed properly. A senior executive from a global pharmaceutical company advises that AI can “identify outliers and harmonize processes.” 70 % of data frontrunners have established ethical guidelines specifically tailored to the use of AI in R&D 61 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Balance building core AI capabilities with strategic partnerships Biopharma organizations should build AI and data platforms internally in order to establish the foundation, as well as to own the infrastructure and knowledge needed to gain long- term strategic advantage from AI. Necessary assets relevant here typically include unified data lakes that integrate omics, clinical, and real-world data; in-house engineering teams focused on data engineering; model validation, and clinical 3 integration; as well as AI governance and ethics frameworks to ensure compliance and trust. Alexandre Malouvier of ICON plc believes that doing certain activities in house is non-negotiable: “Organizations must establish a centralized AI governance structure, ensure ethical, explainable, and responsible AI use, and maintain oversight to prevent unauthorized AI deployments.” That said, organizations can partner with others for acceleration and specialization. Possible areas are working with AI-native firms for generative chemistry and molecule design or clinical trial optimization and patient stratification. Such collaboration gives access to specialized technology and expertise faster than a company could develop them internally. 62 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Build a data- and digital-savvy workforce Biopharma organizations need diverse and specialized skill sets across the R&D value chain, from clinical scientists and researchers to talent skilled in medical and regulatory affairs. In all these domains, critical skills for R&D talent include: • Core technical and data science skills (e.g., ML/DL, data engineering, computational modeling, statistics) • AI application skills (e.g., use case framing, model deployment, automation and robotics, LLMs) • Data privacy and IP management • Model transparency/explainability • Ethical AI (e.g., awareness of bias, reproducibility, sustainability) • Interdisciplinary collaboration and communication (e.g., across scientists, engineers, and data specialists) • Change leadership/mindset 4 • Storytelling (e.g., communicating AI insights to non- technical audiences) • Systems thinking (e.g., understanding how AI fits into the broader R&D ecosystem) • Curiosity and experimentation (e.g., eagerness to test, learn, and iterate rapidly). Interestingly, a slightly higher share of biopharma data frontrunners (59%) require their R&D staff to complete training in AI ethics, with a focus on pharmaceutical applications, compared to data novices (51%). To succeed, an executive from a healthcare manufacturer believes companies need “people who bridge science and digital domains. Team members who can translate between scientific/ process and digital/data domains are incredibly valuable. There is an expectation that the proportion of data architects, engineers, and digital specialists will increase relative to traditional roles.” As AI takes on more routine, manual tasks, biopharma talent can focus on higher-value, creative roles. A senior executive from a global pharmaceutical company says, “AI will elevate scientists' decision-making by automating routine tasks, allowing experts to focus on complex deviations and oversight. Emerging skill requirements include prompting skills and expertise in building and curating data models as increasingly important, alongside collaboration between technical experts and experienced scientists.” 59 % share of data frontrunners who require their R&D staff to complete training in AI ethics 63 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Advocate for organizational data readiness and industry data standards To unlock the full potential of AI in biopharma R&D, organizations must address two critical areas: internal data preparedness and industry-wide data standardization. Ensuring organizations are “data-prepared” internally is critical to realizing the potential of AI. Our research reveals that biopharma data frontrunners are further along in their AI in R&D journeys than data novices. Nearly a quarter (24%) of data frontrunners are actively executing on their strategy and roadmap for AI in R&D, and 58% are finalizing their AI in R&D strategy and are preparing for implementation. This compares to only 1% of data novices who are finalizing their strategy, and no data novices are at the most mature stage of implementing their AI in R&D roadmaps. AI thrives on high-quality, well-structured, and interoperable data. Without robust foundations, even advanced models will underperform. Biopharma R&D generates vast volumes of heterogeneous data – clinical, genomic, imaging – often siloed across systems. Data readiness requires: 5 • Quality and integrity: rigorous processes for cleaning, validating, and curating data • Interoperability: architectures that enable integration across internal platforms and external partners • Rich metadata: contextual tagging for accurate AI interpretation • Scalable infrastructure: cloud or hybrid environments for large-scale data ingestion and processing • Governance and compliance: frameworks to meet regulatory requirements and maintain trust. Equally important is industry-wide data standardization. There is no single standard that covers all of biopharma R&D today. Instead, multiple well-established, domain- specific standards exist, some mature and widely adopted with others still emerging. Shared standards, though, would allow information to flow seamlessly between organizations, CROs, regulators, and AI partners. Thus, by advocating for and contributing to the development of unified data 64 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
standards, organizations can accelerate AI adoption, reduce duplication of effort, and improve the quality and consistency of insights generated. Establishing common ground in data governance will also enhance trust, transparency, and scalability of AI across the industry. As the global head of medical strategy at a Swiss pharmaceutical company puts it, “There is a critical need for industry-wide collaboration for AI to successfully transform R&D and shorten cycle times.” An executive from a healthcare manufacturer describes the long-term possibilities: “Digitalized tech transfer, where it is highly automated, with bots handling tasks currently performed manually by scientists, transforming the CRO business model. Achieving industry-wide data standards, such as a commonly adopted data lake strategy, is essential for realizing AI's full potential.” Organizations that prioritize both internal data readiness and industry collaboration on standards will accelerate AI deployment, unlock predictive and prescriptive insights, and transform R&D efficiency – ultimately reducing cycle times and improving patient outcomes. An executive from a healthcare manufacturer. "Team members who can translate between scientific/process and digital/data domains are incredibly valuable. There is an expectation that the proportion of data architects, engineers, and digital specialists will increase relative to traditional roles.” 65 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
The biopharma industry is undergoing a profound transformation. AI is rapidly emerging as the cornerstone of innovation across the R&D value chain, and in the longer term will drive reinvention and reimagining of the entire industry. It is enabling biopharma organizations to address long-standing, increasingly unsustainable challenges such as productivity declines, slower speed to market, and development inefficiencies. However, realizing AI’s full potential will require more than technological adoption; the technology demands a strategic foundation built on robust data infrastructure, operational readiness, and cultural alignment. This shift will not happen overnight, as many organizations are still building the necessary capabilities. Senior leadership commitment, clear goal setting, and a balanced approach to internal development and external partnerships will be critical. As biopharma organizations mature, AI will evolve from a supportive tool to a strategic partner, helping to steer the next generation of R&D breakthroughs and positioning early adopters to lead in AI-driven transformation. Conclusion 66 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026
Research methodology This publication draws on a multifaceted research program designed to shed light on the strategy and maturity of biopharma organizations in adopting AI in R&D. It included a specially commissioned survey of 500 executives from 250 organizations headquartered across nine leading markets in Asia-Pacific (26%), Europe (30%), and the United States (44%). Over 60% of respondents work for companies with annual revenue greater than $1 billion, with the remainder having revenues between $500 million and $1 billion. Over half (56%) of organizations are in the pharmaceutical sector, 26% are biotechnology companies, with the rest involved in both parts. Nearly half (46%) of executives are in a role that is clinical- and digital/technology-oriented. The survey took place in August and September 2025. The survey also benefits from interviews with 15 senior executives at leading biopharma organizations, as well as extensive desk research. The study findings reflect the views of the interviewees and of the respondents to our online questionnaire for this research and are aimed at providing directional guidance. Please contact one of the Capgemini experts listed at the end of the report to discuss specific implications. 67 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Percentage of biopharma executives by subsector Pharmaceutical Biotechnology Both biotechnology and pharmaceutical 56% 26% 18% Percentage of biopharma executives by job title Director Vice president Executive 32% 47% 5% Percentage of biopharma executives by their organizations' annual revenue $500 mn–$1 bn $1 bn–$5 bn $5 bn–$10 bn $10 bn–$20 bn $20 bn+ 39% 41% 5% 4% 12% 68 Smart bet, only option, or both? Biopharma R&D turns to AI. Capgemini Research Institute 2026 Capgemini Research Institute 2026
Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives. Percentage of biopharma executives by primary role Both clinical-and digital/technology-oriented (e.g., research scientist, analytical chemist, biotechnologist, bioinformatician) Clinical-oriented (e.g., researcher, research coordinator, lab director, QA/QC manager, regulatory affairs specialist) Digital/technology-oriented (e.g., data scientist, IT/systems specialist) 46% 36% 18% Percentage of biopharma executives by function Information technology Innovation Research and pre-clinical trials Clinical trials Data and analytics Digital health/biomarkers Quality (QA/QC) Regulatory affairs Analytical method development (AMD) Manufacturing process development 17% 15% 11% 9% 9% 8% 8% 8% 7% 7% 69 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Percentage of biopharma executives by their organizations' country of headquarters US Germany Japan India UK France Netherlands Switzerland 44% 15% 14% 12% 7% 5% 2% 1% Source: Capgemini Research Institute, Impact of AI on R&D productivity survey, August–September 2025, N=500 pharmaceutical and biotechnology executives 70 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
71 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
References 8. Pfizer, "Insilico enters into research collaboration with Pfizer to explore novel data and AI system for potential therapeutic targets,” January 14, 2020. 9. AstraZeneca, “Data Science & Artificial Intelligence: Unlocking new science insights.” 10. Amgen, “On the global stage: Amgen’s top voices drive conversations on AI and innovation,” May 2025. 11. Sanofi, “Digital and AI.” 12. Isomorphic Labs, “Isomorphic Labs announces Novartis collaboration expansion,” February 2025. 13. Generate:Biomedicines, “Generate:Biomedicines announces multi-target collaboration with Novartis to discover and develop protein therapeutics with generative AI,” September 2024. 14. Microsoft, “Novartis empowers scientists with AI to speed the discovery and development of breakthrough medicines,” November 2021. 15. Clinical Research News, “AstraZeneca scaling up use of generative AI to help reach 2030 ambitions,” March 2025. 16. Fierce Biotech, “AstraZeneca, Algen Biotechnologies pen $555M AI pact for immunology targets,” October 2025. 17. Benevolent, “AstraZeneca starts artificial intelligence collaboration to accelerate drug discovery,” April 2019. 18. AstraZeneca, “Unlocking the future of drug discovery through advances in AI,” September 2025. 19. Roche, “Genentech R&D (gRED).” 20. Recursion, “How a first-of-its-kind ‘microglia map’ from Recursion and Roche and Genentech could unlock new treatments for neuro-degenerative diseases,” October 2025. 21. Sanofi, “AI across the R&D value chain: drug discovery,” July 2025. 22. Novartis, “People and culture from black holes to AI driven drug discovery – collaboration wins the day,” August 2024. 1. Nature Reviews Drug Discovery, “Diagnosing the decline in pharmaceutical R&D efficiency,” March 1, 2012. 2. Pharmaceutical Research and Manufacturers of America (PhRMA), “Progress toward new medicines and vaccines,” October 2024. 3. Science Direct, “Why 90% of clinical drug development fails and how to improve it?” July 2022. 4. Science Direct, “Why 90% of clinical drug development fails and how to improve it?” July 2022. 5. US Food and Drug Administration, “FDA launches agency- wide AI tool to optimize performance for the American people,” June 2, 2025. 6. European Medicines Agency, “Data and AI in medicines regulation,” July 2025. 7. Exscientia and Sanofi establish strategic research collaboration to develop AI-driven pipeline of precision- engineered medicines,” January 7, 2022. 72 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
23. AstraZeneca, “Data science and artificial intelligence: Unlocking new science insights.” 24. Gilead, “Gilead and Genesis Therapeutics announce strategic collaboration to discover and develop novel therapies,” September 2024. 25. Pfizer, “Artificial intelligence: on a mission to make clinical drug development faster and smarter,” 2025. 26. Fierce Biotech, “Novartis makes investment in AI firm Yseop for clinical trial writing,” December 2023. 27. Clinical Research News, “AstraZeneca scaling up use of generative AI to help reach 2030 ambitions,” March 2024. 28. Novartis, “Novartis' commitment to the ethical and responsible use of artificial intelligence (AI) systems.” 29. Sanofi, “Digital and AI.” 30. AWS, “Key takeaways from Gilead’s innovation talk at re:Invent 2023,” March 2024. 31. Sanofi, “AI across the R&D value chain: drug discovery,” July 2025. 32. Netguru, “Speeding up Merck’s process from 6 months to 6 hours with an AI R&D assistant,” 2024. 33. Bristol Myers Squibb, “Science firsthand: Predicting new possibilities in drug discovery,” February 2025. 34. Bristol Myers Squibb, “Four (and more) ways AI is powering innovation and efficiency in our workplace,” November 2024. 35. Clinical Leader, “How BMS uses AI to improve study design and reduce costs,” January 2023. 36. Yseop case study, “The road to an automated clinical document landscape at Eli Lilly,” June 13, 2023. 37. Lilly, “Lilly launches TuneLab platform to give biotechnology companies access to AI-enabled drug discovery models built through over $1 billion in research investment,” September 2025. 38. Clinical Research News, “AstraZeneca scaling up use of generative AI to help reach 2030 ambitions,” March 2025. 39. Clinical Research News, “AstraZeneca scaling up use of generative AI to help reach 2030 ambitions,” March 2025. 73 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Authors Thorsten Rall Executive Vice President, Global Life Sciences Industry Lead, Capgemini thorsten.rall@capgemini.com Brian Eden Vice President, Global Life Sciences, Capgemini brian.eden@capgemini.com Justin Melnick Director, Global Life Sciences Incubator, Capgemini justin.melnick@capgemini.com With more than 15 years of industry experience, Thorsten is responsible for globally setting and executing Capgemini’s strategy in life sciences as well as working with key clients on their business transformation toward leveraging data, artificial intelligence, and technology at scale. He has always been driven by the goal to achieve better care for more patients more quickly around the globe. Combining his longstanding work experience in professional services and in the pharma industry, Thorsten brings together his deep understanding of industry, business strategy, and hands-on execution to drive tangible value generation across the life science value chain. He drives value at enterprise level and within specific focus areas including clinical development and marketing and sales. Prior to Capgemini, he held various senior positions at Novartis. As Senior Vice President of Digital Transformation and Innovation, he led the digital function, developed and implemented high-risk, high-return strategic innovation initiatives, and was responsible for business development and partnerships in the digital space. Brian has three decades of operations transformation, management consulting, and leadership experience. He has developed and led global Lean and Digital programs, consistently recruiting and developing a diverse array of talent, and has led global business integrations, executed a range of process and product programs, and directly supervised operations of technical processes. Brian’s experience extends to government, public and private sectors, and to the pharmaceutical, medical device, energy, consumer products, and food processing industries. He is a Certified Master Black Belt, Change Management Expert, Lean Expert, GE trained executive, Veteran US Naval officer, and certified Nuclear Engineer Justin is an experienced leader at the intersection of computational and biological sciences, focused on applying emerging technologies to life sciences R&D. With over 12 years of experience, he has successfully led cross- disciplinary teams, driving value through product management, generative AI strategies, and AI-powered drug development. Justin’s recent work includes projects on mRNA vaccine data, electronic lab notebooks, and portfolio management for early-stage R&D. Recognized among the top 100 solution architects at a leading AI advisory firm, Justin combines deep expertise with a collaborative approach to solving complex R&D challenges. Passionate about using technology for real-world impact, he is dedicated to helping life sciences organizations optimize their innovation processes and improve patient outcomes. Meet the experts 74 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Marisa Slatter Director, Capgemini Research Institute marisa.slatter@capgemini.com Jerome Buvat Head of Capgemini Research Institute jerome.buvat@capgemini.com Authors Meet the Capgemini Research Institute Harshada Suresh Sambare Manager, Capgemini Research Institute harshada.suresh-sambare@ capgemini.com The contributors would also like to thank Frédéric Burger, Martin Brock, Sheetal Chawla, Symon Cotton, Etienne Grass, Deepak Gupta, Raghunandan Hanumanthu, Caren Hochheimer, Malina Keutel, Ivana Knyght, Gert-Jan Schoppert, Geoff McCleary, Parvin Moyasari, Nirlipta Panda, Mark Roberts, and Richard Traherne for their contributions to the research. 75 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
For more information, please contact: Global Thorsten Rall thorsten.rall@capgemini.com United Kingdom Justin Melnick justin.melnick@capgemini.com North America Brian Eden brian.eden@capgemini.com France David Ghesquieres david.ghesquieres@capgemini.com 76 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
The Resonance AI framework by Capgemini Waves of AI value The Resonance AI framework by Capgemini provides a sequential approach to the conceptualization, structuring, and implementation of successful AI-driven transformation. It helps business leaders realize AI’s potential and achieve market leadership regardless of the industry. Anchored in transformation strategy, the framework helps integrate operations and culture while accelerating AI value creation – to both transform today and build for tomorrow. A d o p t A d a p t ACCESS ADOPT ADAPT HUMAN-AI-CHEMISTRY Adoption to collaboration AI ESSENTIALS Tech to data AI READINESS Enablers to guardrails Transform now Operational efficiency hyper-personalized experience Build tomorrow Business (re)invention next frontier innovation 77 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Delivering AI at scale for life sciences What it takes to do truly data-driven R&D of the future Human-centricity Product-led-software Data foundation Hybrid- AI-systems Findable Accessible Interoperable Reusable Applied (Gen)AI DDR&D the future F A I R Application of mathematical modelling, quantum computing and machine learning to simulate complex molecular interactions with unprecedented accuracy, providing insights that classical technologies cannot achieve improving speed to market and probability of candidate success Specialized GenAI toolsets that would augment scientific teams by being: • Knowledge-based: understanding the scientific context • Specialized reasoning: generates new formulations and even suggest how to produce them • Verified: leverages scientific knowledge models to validate outputs against norms and standards Combining different types of AI and non-AI technologies to create tools that empower research teams to use data to develop new drugs and therapies, improve clinical trials and bring them to market in a fast and sustainable manner. Your data foundation covers People, Processes and Technology . It supports building a common understanding that data is a business asset and defining a structured framework to support research goals where good data practices permeate the entire organization. Product-led software drives business model innovation by leveraging existing assets - human energy and data - to unlock new capabilities. It transforms experience into products through a data-first, digital native approach, augmented by digital (GenAI) coworkers to accelerate feedback loops and ensure continuous value delivery. 78 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Harnessing the value of AI Connected health Rise of agentic AI Lab of the future Engineering biology AI in action More Capgemini Research Institute publications 79 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Receive copies of our reports by scanning the QR code or visiting https://www.capgemini.com/capgemini-research-institute-subscription/ Subscribe to latest research from the Capgemini Research Institute 80 Capgemini Research I nstitute 2026 S m art bet, only option, or both? B iopharma R&D turns to A I .
Notes 81 Capgemini Research Institute 2026 Smart bet, only option, or both? Biopharma R&D turns to AI.
Smart bet, only option, or both? Biopharma R&D turns to AI. Copyright © 2026 Capgemini. All rights reserved. About Capgemini Capgemini is an AI-powered global business and technology transformation partner, delivering tangible business value. We imagine the future of organizations and make it real with AI, technology and people. With our strong heritage of nearly 60 years, we are a responsible and diverse group of 420,000 team members in more than 50 countries. We deliver end-to- end services and solutions with our deep industry expertise and strong partner ecosystem, leveraging our capabilities across strategy, technology, design, engineering and business operations. The Group reported 2024 global revenues of €22.1 billion. Make it real | www.capgemini.com