dot Be the first to see our latest product releases—virtually—at Redis Released: Worldwide.

Register now

The Data Economy: How Novartis is Becoming a Data Science Company with AI Innovations

The Data Economy is a video podcast series about leaders who use data to make positive impacts on their business, customers, and the world. To see all current episodes, explore the podcast episodes library below.


“The decisions that we make based on the models will affect patients’ lives.”

Bülent Kiziltan, Ph.D., Head of Causal and Predictive Analytics at Novartis’ AI and innovation lab, shared this important point during his recent interview on The Data Economy CXO podcast, presented by Redis and hosted by Michael Krigsman of CXOTalk. A big part of the AI lab’s mission is to improve efficiencies in bringing new drugs to market, and Kiziltan stresses how data, analytics, industry partnerships, and talent from diverse backgrounds are core to their strategy. Kiziltan discloses that the goal is to augment the drug development and discovery process, and they are partnering, not replacing the lab because the stakes are so high. 

Here’s how Kiziltan articulates the goal: “The cost goes up linearly with the increased time spent on a particular drug development and discovery process. We can shorten the period by creating additional evidence using the data science methodologies that we are developing, and also target specific patients and cohorts that will be more suitable for specific studies.”

Kiziltan says that Novartis started its transformation into a data science company several years ago. The AI innovation lab is over a year old, and in addition to the causal and predictive analytics discipline that he leads, there are natural language processing and image analytics pillars. The three groups have access to 20 petabytes of data, including patient data from 2 million people, images, scans, and a library of close to 2 million compounds.  

And how is all this data managed? Novartis technology group adopted a hybrid-cloud strategy to meet their needs and partnered with Microsoft Research on the technology and implementation. The AI lab is one consumer of this technology. In the domains of healthcare and biotech, “the more data, the better,” says Kiziltan. “But the amount of data is a constant problem because it takes a long time to gather and clean. Providing access across different business units can be challenging because of the stringent regulation.”

https://www.youtube.com/embed/fm3XA-uGwqE

Advice for CIOs on transforming into a data science company

Novartis is a Swiss multinational pharmaceutical company with over $48 billion in revenue and 100,000 employees. Now that it has transformed into a data science company, Novartis is reimagining its enterprise for both long-term opportunities and short-term gains. 

In the episode, Kiziltan shares a key lesson for CIOs, Chief Data Officers, and AI/analytics leaders: “To sustain the value creation that comes from data science and AI, a smart, strategic roadmap is to invest both in R&D-driven, exploratory core capabilities as well as building strategies for real-world impacts that can be executed immediately.”

Enterprise CIOs leading digital transformations should note how Kiziltan guides investments toward long-term competitive advantages while reducing delivery costs. Novartis’ investment in its AI lab is a business opportunity to improve resiliency and agility. 

In the podcast, Kiziltan shares some of the operational and data challenges his team faces:

  • Growing the infrastructure to make data directly accessible and ready to analyze
  • Handling sparse data and effectively and efficiently extracting new insights from limited information 
  • Combining data from different streams (in the case of Novartis, data from different labs, vendors, and trials)
  • Working with longitudinal data, which is data recorded at different times, and various types of structured and unstructured data
  • Automating operational processes using a mix of supervised, unsupervised, and semi-supervised methodologies
  • Learning the latest data science techniques to address data biases, including biases inherent to the data and biases from the selected methodologies
  • Partnering with social scientists and academics on the development of ethical AI, which is an important consideration when applying machine learning in healthcare

CIOs and their teams likely face some of these challenges, and Kiziltan suggests following a single strategic approach to collaborating and finding solutions.

CIOs must focus on hiring diverse data science talent

While a variety of technical skills are required to solve these kinds of problems, Kiziltan recognizes that hiring a diverse team with strong problem-solving skills is also important because data science is changing so rapidly.

“We have been attracting talent that can bring in their diverse backgrounds into our operations,” he says. “We hire from very different domains, including physics, mathematics, bioinformatics, and chemistry. Even people with a social sciences background have built some analytical skills that are contributing to our ongoing efforts. I focus on curiosity and the potential to learn.”

In fact, Kiziltan himself comes from an astrophysics background and has applied mathematics from this discipline to his work.  He says, “Innovation can happen in two ways: One way to innovate is to build and develop new methodologies, but the other means of innovating is to apply methodologies to new problems.”Please tune in to the podcast to hear more of Kiziltan’s insights on AI and the transformation of Novartis into a data science company.


Watch more episodes of The Data Economy podcast.