In this episode of Eureka!, a McKinsey podcast on innovation in life sciences R&D, hosts Navraj Nagra and Alex Devereson speak to Marcus Schindler, executive vice president for R&D and chief scientific officer at Novo Nordisk. They discuss how Novo Nordisk is expanding its external innovation capabilities and moving into new therapeutic areas. They also explore his efforts to embed AI throughout the R&D organization, establish Novo Nordisk as a leader of AI ecosystems (in Boston and beyond), and eventually advance from treating diseases to curing them, with help from AI.
Leading through a period of fast growth and expansion
Navraj Nagra: What have you been focusing on at Novo Nordisk?
Marcus Schindler: I was hired to create an external innovation capability. At the time, Novo Nordisk had few, if any, partnerships. Our CEO, Lars Fruergaard Jørgensen, felt that, although Novo Nordisk was a successful player in niche diseases—diabetes, in particular—he wanted to take the company to the next level. We needed to open up and become more curious about what others were inventing and how they worked.
In my previous roles, I worked a lot with the external world, so I was well-versed in it and interested in bringing my expertise to Novo Nordisk. My mission was twofold: bring in external innovation and help the company expand into therapeutic areas beyond diabetes, which we are doing.
Navraj Nagra: Novo Nordisk has grown very quickly in the last couple of years. How are you and other leaders managing the change and staying focused?
Marcus Schindler: First and foremost, we have clarity on who we are, who we want to be, and what we want to contribute to the world of pharmaceuticals and beyond. We are trying to solve complex, health-related societal problems with a focus on a very limited number of diseases. That’s the way I articulate our work not only externally but also internally. The growth is a secondary benefit. We’ve had incredible financial success because we’ve taken obesity—a massive unmet medical need for decades—and turned it into a treatable condition. That hasn’t happened very often in our lifetimes. We have achieved this without lowering our scientific standards. There are many opportunities we could pursue, but that doesn’t mean we should. It takes a lot of discipline and a lot of saying no.
Additionally, within R&D, we’ve set a limit of 10 percent headcount growth per year to ensure we successfully integrate new hires to get the best from them and help them build strong networks. That takes time and effort for the ones joining and for those already in the organization. Giving people that time is crucial.
Last, we are an organization that is majority-owned by a foundation,1 so we have a long-term view. We have incredibly strong values, including what we call the triple bottom line. We are responsible for several things: our patients, sustainability, the environment, and financial performance, and they are baked into all our decision making. At the end of the day, we’re staying focused, disciplined, and true to ourselves.
Embedding AI tools and capabilities throughout R&D
Alex Devereson: To what degree have AI and machine learning become part of R&D at Novo Nordisk?
Marcus Schindler: The big wave of automation and digitalization at Novo Nordisk is very recent. When I joined in 2018, we were only just then digitalizing our lab books. We started with the basics to set up a genetics unit and then developed analytics capabilities and added computing power. That allowed us to consider what we’d like to do with digitalization and machine learning. Handheld, unstructured efforts were growing in all parts of the organization, solving specific needs of specific users.
A few years ago, we decided to create a digital strategy that was fit for our purpose. We weren’t even talking about generative AI [gen AI] at that point. We created an organization that works with digital signals and innovation. Through that organization, we hired people with competencies we lacked. It really became our driving force and brought a lot of excitement and energy. But it also had challenges defining itself: was it a service organization or a driver of innovation? The answer is both. It’s been a journey of constant learning, but when I take stock of our capabilities and partnerships today, we’ve made a lot of progress.
Personally, I’ve needed to stay humble and open to learning from scratch—for example, by reading Artificial Intelligence for Dummies. It’s important for senior leaders to admit we don’t know everything but are willing to learn and listen and to surround ourselves with people who are way smarter than us.
Alex Devereson: How did you develop your team’s technical capabilities and promote AI- and digital-enabled ways of working?
Marcus Schindler: In our industry and beyond, AI and gen AI are tremendous opportunities, but as with many other revolutionary technologies—recombinant technology and genetics in the early 2000s, for example—they are difficult to fully implement and take time to land well. We pressure-tested the initial promise, endured some disappointments, and overestimated. Over time, we are learning how to create truly agile organizations rather than repeat old mistakes and build new silos.
We started in R&D with a digital science unit of 250 people—we now have about 2,500 people globally—but we expected everybody in the organization to become good at both. If you’re a good chemist or biologist, you also need to understand gen AI and master digital technologies.
Where human intelligence will stay relevant is in formulating the biological medical questions that we’re trying to solve. That’s much harder than we ever thought it would be. From the outset, we were keen to create a culture focused on a unified way of working rather than on organizations. Whatever we are doing, AI will penetrate it.
Many AI tools are being quickly adopted. ChatGPT reached 100 million users much faster than any previous platform. We’re adopting our own version of tools like ChatGPT, but we also have aspirations. One of ours is to do drug discovery without animal testing. We can’t achieve our aspirations without the help of AI.
In another example, it currently takes us 1,500 days, on average, from the time a drug enters the pipeline to human testing and up to ten years to market. Post-COVID-19, we have an aspiration of 500 days to clinical testing. This isn’t possible with traditional experiments, but you can iterate faster with AI.
Unlocking new opportunities through AI for exploration of disease
Alex Devereson: What barriers have you overcome with AI and machine learning, and where have these technologies had the biggest impact to date?
Marcus Schindler: This is a classic case of change management. Something novel appears—in this case, digitalization and gen AI—that we can’t directly control because it is still largely in the hands of the tech industry. So we’re watching this space and trying to understand what it means for us. With its potential for efficiency gains, it looks like a threat to businesses and individuals because it could eliminate some roles in the future. Then the question is—how can humans evolve through learning and adaptation to perform roles only we can do and leave machines to do what they do even better? That doesn’t happen overnight. We’re heavily investing in training and education, some of which is mandatory, but we could offer even more because increasing people’s understanding of these technologies eases their anxiety, which accelerates adoption.
It doesn’t end there, though, because the machine still has to prove it can do something useful. When biotech companies started to put the first molecules entirely based on AI machine learning into clinics, some failed, and people said, “I told you so; we need to go back.” But these are the first ones out of the block, and they will learn. We need to continually manage expectations.
Once we decided we were ready to own or codrive AI solutions, we created some really cool collaborations and partnerships. For example, we’re working with Microsoft to use natural language processing to more deeply understand diseases such as atherosclerosis. We’re also collaborating with some other companies on more-specific problems.
I think the machine will simply create a new reality. It will not be a human reality that we grasp; it will be the machine-learned reality. Maybe we’ll never understand how the machine arrives at its conclusions, but it will get to the true essence of disease. Think about facial recognition: the machine is now better and more accurate than we are at recognizing any animal. If you apply this to disease, it will be deep and revolutionary, in my mind.
Harnessing competitive advantages in ecosystem engagement
Navraj Nagra: How will Novo Nordisk sustain its competitive edge in this space?
Marcus Schindler: First, Novo Nordisk stands for quality. The quality of our data and our quality mechanisms are very good. We’re probably known for running the most and the largest cardiovascular outcome trials in our field. My colleagues in late-stage development do this with extreme precision and deliver amazing results. With this comes an amazing data set that often only we can access. Now we’re building a data foundation with all our clinical and preclinical trial data—not just a random collection of a turbid soup, but something high quality that we can truly interrogate. Building this data foundation took us a while, but it was absolutely worth the effort.
Second, as a company owned by a foundation, we have a long-term view. We’re not here for quarterly results, although those are of course important. We’re working on some bigger problems that will take years or maybe even a decade to come to fruition. One reason I joined Novo Nordisk was the hope of leaving a legacy of work for others to complete.
In short, unless we become complacent, and there’s no evidence of that, our high-quality data coupled with how we work and who we are as a company will help us maintain our edge.
Navraj Nagra: What are the benefits of engaging with the AI ecosystem, and what have you learned?
Marcus Schindler: Partnerships have given us a playground where we can foster our sense of curiosity and openness. The Microsoft collaboration was one of our first, but we have other partnerships, and each one is different. It’s about cocreation—how we can create something that doesn’t exist that neither of us alone could make.
Jointly having a manifesto informs the partnership behaviors. Sometimes we face harsh realities, things go wrong, and emotions run high. It’s good to take a step back and remind ourselves why we are here together in the first place. It helps us practice behaviors of curiosity, openness, and generosity and allows everyone to contribute and even sometimes make mistakes.
Navraj Nagra: In 2023, Novo began consolidating its R&D operations in Boston, an established biological and tech ecosystem. What are you hoping to accomplish with this move?
Marcus Schindler: We cannot afford to not be present in Boston. Despite all the advances in digital communication, physical presence in ecosystems and communities still makes perfect sense because it supports human interactions and builds trust and mutual understanding. It’s a source of inspiration and potential partnerships for work on our diseases with the technologies we’re using and, more importantly, with those we should be using in the future.
