##  [Smarter decisions at scale: How Novartis is using AI to advance R&D](/stories/smarter-decisions-scale-how-novartis-using-ai-advance-rd "Smarter decisions at scale: How Novartis is using AI to advance R&D") 

 [Discovery](/tags/category/discovery)

 

 

#  Smarter decisions at scale: How Novartis is using AI to advance R&D 

Novartis R&D is harnessing digital technologies to make smart decisions faster, aiming to reduce the time it takes to bring new options to patients



 

 Jul 28, 2026 

Across Novartis, accelerating practices that turn hypotheses into evidence and molecules into medicines is a priority. In research and development, our teams are taking advantage of rapid advances in digital technologies to cut the time it takes to discover and develop drugs for patients while further improving on quality, safety, and the patient experience.

These technologies, many of them powered by artificial intelligence (AI), are increasingly helping teams process vast amounts of information faster and answer critical questions earlier, with greater confidence. These questions span the entire research and development spectrum and include:

- Which biological targets look most promising?
- Which molecules might work on those targets with the fewest side-effects? And,
- Which clinical trial design will provide the insights we need, in the right patient population, as efficiently as possible?

Our Novartis R&D leaders, **Fiona Marshall**, President of Biomedical Research, and **Shreeram Aradhye**, President of Development and Chief Medical Officer, have defined a clear strategy for how these technological advances should be integrated from end-to-end, across the R&D continuum. The goal is not technology for its own sake, but a practical, scalable approach that helps teams move from data to decisions more efficiently.

### What does a digital strategy built for speed, scale, and impact look like in R&D?

Novartis R&D has focused on embedding technologies that speed development and improve the likelihood that promising scientific ideas become medicines in a field where less than 10 percent of drug candidates historically reach the market.1

Our teams apply AI and digital technologies according to a clear set of guiding principles. Applications should be:

- **Purpose-driven**, helping to address specific scientific and operational challenges, such as shortening iterative cycle times during drug discovery and clinical testing, improving quality, or surfacing safety signals earlier.
- **Smart and scalable**, designed to be reused and improved over time and across therapeutic areas, teams, and geographies—not just one-off pilots.
- **Human-centered and responsible**, built to augment the judgment and expertise of our teams, with safety, quality, and privacy controls embedded throughout to keep patient outcomes and scientific integrity front and center.

These efforts are shifting how teams work so they’re spending less time sifting through data and more time on creative iteration and decision-making. This matters both to patients waiting for new therapies and to the teams responsible for advancing them. Fewer iterations and shorter cycle times enable faster progress through the R&D pipeline, with decisions increasingly informed by a more complete view of the data.

> By making each AI incremental intervention valuable on its own and collectively transformative, we aim to accelerate the evidence generation that turns molecules into medicines so that patients can receive better treatments faster.
> 
> \- **Shreeram Aradhye**, M.D.



 ![Shreeram Aradhye - President, Development and Chief Medical Officer](/sites/novartis_com/files/styles/crop_freeform/public/2026-06/shreeram-aradhye-media-library-2.jpg.webp?itok=zj_vdQVj "Shreeram Aradhye - President, Development and Chief Medical Officer")

 







 ![Fiona H. Marshall - President, Biomedical Research](/sites/novartis_com/files/styles/crop_freeform/public/2026-06/fiona-marshall-media-library-1.jpg.webp?itok=8xUkbn-w "Fiona H. Marshall - President, Biomedical Research")

 



> Drug discovery is a resource-intensive, time-demanding process in which success is far from guaranteed… AI shows great potential to cut through the noise, to help us make insights faster and inform smart decisions.
> 
> \- **Fiona Marshall**, Ph.D.







### Our R&D roadmap for digital impact

R&D is rooted in complexity. Biology rarely offers clean signals, and modern medicine generates more data than any one person or team can interpret unaided. Today’s digital tools are helping our teams connect the dots across experiments, trials, scientific literature, and real-world evidence, so they can spend less time crunching numbers and more time making informed decisions.

This is playing out in focused initiatives across Novartis R&D:

**Target identification and validation:** AI models can integrate evidence from human genetic and multi-omic datasets, imaging, clinical trial results, and decades of published research to surface previously underappreciated disease drivers and strengthen confidence in the targets teams choose to pursue. This is empowering teams to move beyond extensively studied targets to focus on new avenues, backed by converging lines of evidence. Combined with the right lab experiments, this technological edge can prompt a breakthrough that would have taken many years using traditional approaches alone.

**Generative chemistry:** Historically, drug hunters have often had to discover or design new medicines through traditional high-throughput screening of millions of molecules and by optimizing chemical structures through countless iterative cycles. AI is changing this. With generative approaches, teams can propose new structures informed by protein information and prior chemistry, explore large, molecular-design spaces digitally, and optimize for multiple chemical properties in parallel, helping scientists move more quickly from “hypothesis” to “drug candidate,” with fewer iterations.

**Preclinical safety:** By learning from decades of preclinical and clinical data, AI can help flag potential safety concerns earlier, supporting teams in deciding what to advance, what to refine and what to stop—decisions that can be a challenge in biopharma. Having this clarity at the earliest point possible saves time, reduces risk, and ultimately helps ensure that the drug candidates that advance through our pipeline are the ones with the best chance of helping patients.

> It’s important that we responsibly leverage every tool available to speed up the drug development process so that we can deliver new options for patients waiting for the next breakthrough.
> 
> \- **Fiona Marshall**, Ph.D.

> We’re implementing these technologies in a way that ensures that we’re making better-informed, better-quality decisions that can translate into an improved rate of success.
> 
> \- **Shreeram Aradhye**, M.D.

**Clinical trial design:** AI-enabled tools such as Generative AI, paired with knowledge graphs, can rapidly summarize what has worked—and what hasn’t—across prior studies and real-world evidence. That helps teams draft and refine protocols faster, pressure-test key assumptions earlier, and design trials that are better aligned to the scientific question and patient population that our R&D teams are focused on. Simulations using trial computational twins can also help predict the operational implications of clinical design elements, enabling swift testing of what-if scenarios, making it simpler to plan clinical trials that operate as intended, on shorter timelines.

**Site selection and recruitment:** By combining clinical, operational, and real-world datasets, AI can help identify the sites and geographies most likely to reach the right patients for each study. Better targeting supports faster and more representative enrollment and reduces trial delays.

**Operational precision in clinical trials:** Novartis teams are using advanced, AI-enhanced analytics to surface early risk signals across supply, monitoring, and recruitment, and to recommend “next best actions” when risks are identified to keep studies on track and avoid major delays. This approach supports more proactive risk management, faster course corrections, and more predictable execution of studies.

**Clinical and regulatory documentation:** In R&D, documentation must be thorough, accurate, and consistent. It can be a time-consuming task. Generative AI can create well-structured first drafts using approved sources, which Novartis experts then review and refine. That can reduce time spent on repetitive formatting and synthesis and free teams to focus on the scientific and medical judgment that only experienced professionals can provide.

### How are we scaling AI responsibly, with people in mind?

To make AI as impactful as possible, Novartis R&D relies on both internal and external innovation. We form strategic collaborations with leading AI companies in areas where their AI platforms and knowledge complement Novartis healthcare expertise. At the same time, we leverage our extensive background in data and digital technologies to develop fit-for-purpose tools internally when off-the-shelf solutions don’t meet the needs of regulated, evidence-driven science.

But no matter the tool, we don’t lose sight of the human creativity and scientific insight that is so essential to making medicines. Novartis R&D’s AI strategy is human-centered, augmenting our teams’ expertise. Ultimately, the ambition is simple: help scientists and clinicians spend less time searching, reformatting and second-guessing, and more time making high-quality calls that move programs forward.

This fusion of computational power and human, scientific, and medical know-how is a partnership designed to accelerate the pace of R&D and empower our workforce to make smart decisions based on an unprecedented wealth of data, to bring impactful therapies to patients as efficiently as possible.

> While the contributions of AI will be invaluable, medicine will always be a human endeavor.
> 
> \- **Fiona Marshall**, Ph.D.

> We believe that enabling all of the great people that we have here at Novartis with AI tools is fundamentally going to evolve our ways of working so that we bring drugs from discovery to patients faster and with a higher likelihood of success.
> 
> \- **Shreeram Aradhye**, M.D.



 



###  To hear more from our R&D leaders on this topic: 

1. Read Fiona Marshall’s articles on AI in Biomedical Research:
    - [Here’s how AI is reshaping drug discovery](https://www.weforum.org/stories/2026/01/how-ai-is-reshaping-drug-discovery/), World Economic Forum “Forum Stories”
    - [X-rays to AlphaFold: the future structure of drug discovery](https://www.linkedin.com/pulse/x-rays-alphafold-future-structure-drug-discovery-fiona-h-marshall-0rboe/), Fiona Marshall on LinkedIn
2. Watch Shreeram Aradhye’s videos and interviews on how AI is being applied in Development:
    - [Perspectives from the frontline of AI: Lesson 08](https://faculty.ai/lesson-08-novartis), Faculty AI
    - [Harnessing the power of AI in drug development](https://www.linkedin.com/posts/shreeram-aradhye-80a3b95_in-my-mind-it-is-clear-that-artificial-intelligence-ugcPost-7198617380745314305-fYhx/), in conversation with John Gibson, Chief Commercial Officer, Faculty AI



 

 





 

**References:**

1. Zhou, Y., Zhang, Y., Xu, H. et al. Dynamic clinical trial success rates for drugs in the 21st century. Nat Commun 16, 9537 (2025). [https://doi.org/10.1038/s41467-025-64552-2](https://www.nature.com/articles/s41467-025-64552-2)