Medical Moonshot: How Novartis and Microsoft Are Using AI to Reimagine Medicine – Cloud Wars

While pharmaceutical companies have traditionally had to devote $2.5 billion and 12 years of to bring a single therapy to market, the stark reality is that only 1 out of 10,000 compounds makes it through clinical trials and into the hands of patients.

Determined to take these challenges head on, Novartis and Microsoft have created an AI Innovation Lab with the audacious goal of shattering those barriers by reimagining not just medicine but also daring to approach R&D in part as a computational challenge.

At the heart of that visionary collaborative effort is the fusion of expertise in deep scientific know-how from Novartis, with AI and data management expertise from Microsoft. On top of all that world-class domain expertise, leaders within both companies say, the other essential ingredient is culture: can the Novartis and Microsoft teams transcend the traditional dynamic of tech vendor and business customer to achieve outcomes that neither company could ever accomplish individually?

The partnership with Microsoft is one that we at Novartis value and cherish, said Novartis chief digital officer Bertrand Bodson, who joined the company 2-1/2 years ago after stints in online retail, digital music, entrepreneurship and Amazon.

It brings together great minds from science, technology and data to find the best ways to build platforms, explore data, and improve the speed at which we operate.

A part of my job I most enjoy is that when I see and talk to the teams doing this incredible work, I cant tell who is from Novartis and who is from Microsoft. And when you achieve that sort of focus and trust, great things can happen, said the relentlessly upbeat and fast-talking Bodson.

Among the great things Novartis hopes to achieve:

A key component to becoming this leading medicines company powered by data and digital is an initiative called data42, which will allow Novartis to derive better insights from its more than 2 million patient-years of clinical-trial data that will serve as the fuel for the AI systems Novartis is creating in collaboration with Microsoft, Bodson said.

In the past, these datasets were not unified so could not be analyzed together as a whole. But now that were doing this, we believe we will be able to build scalable products on the data42 platform, allowing our scientists to be able to gain more insights and deeper insights than has ever before, Bodson said.

The new data architecture developed by Microsoft and Novartis makes it possible for Novartis scientists to analyze and work with vast and aggregated datasets rather than limited and fragmented results that have simply been too small or too limited to drive high-scale innovation at the speed Novartis desires.

We are now able to bring to bear the predictive powers of AI against this massive reservoir of data, Bodson said. For the first time, we can ask, If were not recruiting patients fast enough in a certain geographic area, how should we rethink patient-recruitment so that we reach our goals?

Or we can use those predictive powers to enhance the molecules were creating and work with because we can now more precisely probe biological systems. And that can lead to breakthroughs in vital areas such as smart dosing.

At the center of those efforts is the AI Innovation Lab established in September 2019 by Novartis and Microsoft.

Its charter is to help the market-leading pharmaceutical firm dramatically ramp up its capabilities around data science and AI with a specific focus on healthcare and life sciences.

As we use AI to explore exciting new fields such as generative chemistry by simulating the properties and behaviors of molecules in ways that yield insights faster than ever before, thats where the real magic happens because we are deeply fusing Microsofts expertise in AI and data science with our own deep expertise in biology and life sciences, Bodson said.

Another opportunity for that type of magic is turning those one-off achievements into documented and repeatable processes that can be leveraged across the company and across the world to drive the type of scale that can dramatically reduce the barriers of drug hunting.

An article on the Novartis website about data42 offers this perspective from Pascal Bouquet, whos the Technology Lead for the platform: Of course, we dont know yet what we are going to find when we are using this new data and digital technology. But we firmly believe we will be able to find insights that are not possible today.

We are convinced that we can find nuggets that we have not seen so far and that, in the long run, we can even completely design and discover new drugs based purely on data, Bouquet says in the article (emphasis added).

For that to happen, Bodson says, Novartis must embrace a data culture from one end of the company to the otherwithout that type of environment, it will be impossible to develop the scale needed to shrink the cost and development time of new medicines.

Building a data culture is a very big part of my jobprobably 50% of it, said Bodson. We have to treat data as a high-value asset that belongs to everyone at the company, and is not isolated here or there for only a few people to be able to use. But I must tell you, all of that is easier said than done.

To help foster that essential data culture, Bodson said, the Novartis/Microsoft team running the AI Innovation Lab is taking a decidedly outside-in approach to foster awareness, trust and collaboration from all segments of the vast scientific community within Novartis.

We dedicate time to better understanding the needs and requirements of our teams. What are your top challenges? Where can we help? How can we help you get the insights you need? The team understands that we want them to be active and to push into new areasif they fail, thats fine, but dont be afraid to take chances based on the data you have.

Those efforts from the AI Innovation Lab generally coalesce around two areas: AI exploration, which involves tackling some of the hardest challenges within life sciences, very specific projects, starting with generative chemistry and optimizing cell and gene therapies at scale. The second area is AI empowerment, which Bodson said is a Microsoft term centered on enabling every person to harness AI in their own special and custom ways to reason over diverse information and unlock valuable new insights as they simultaneously reimagine ways to achieve those results more rapidly and at lower cost.

From his role as chief digital officer, Bodson helps align all of those efforts against four strategic pillars for Novartis:

While were confident in our abilities at Novartis to make big contributions, the fields of medicine and biology are simply too big for any one company to handle. Its too big for us to do it alone, Bodson said.

So one of the reasons we were so excited about working with Microsoft is they have the same outlook and the same openness about partnerships. And their expertise in building platforms, in AI, and in helping big global customers like Novartis build with agility is incredibly important to us. We know theyve invested massively to create that expertise, Bodson said.

Its important because some of the big technology players have tried to move into the healthcare business on their own. But one thing I learned very quickly at Novartis is that biology is humbling and amazingly complex.

Were taking on some of the biggest healthcare challenges out there, and for that we need this powerful pairing of science and tech as theyre much, much more than simply computational problems.