Leveraging Data to Add Value with Marc Paradis of Northwell Holdings

This is part of our weekly executive insights series where Elion CEO Bobby Guelich speaks with healthcare leaders about their tech priorities and learnings. For more, become a member and sign up for our email here.

Role: Vice President of Data Strategy

Organization: Northwell Holdings, a for-profit investment arm of Northwell Health

Can you give an overview of your role and what you’re focused on at Northwell?

As the VP of Data Strategy for Northwell Holdings, my role is essentially fourfold:

• Evaluating companies and opportunities in the Northwell Holdings portfolio through the lens of data science, machine learning, AI, data partnerships, and data strategy.

• Helping to shape Northwell’s broader data strategy.

• Leading a small team of data engineers who have built a data lake house covering over 7 million patients.

• Contributing to the leadership team of Northwell’s newly established AI Center of Excellence.

Let’s start with data strategy—what does this mean to you and how do you approach it at Northwell?

Data strategy can mean a couple different things. There’s traditional IT-focused data strategy concentrating on integrating systems, building enterprise data resources, and governance, and then there’s leveraging data for innovation, improving operations, and generating revenue, which is where I focus. Along those lines, one of the things I’m thinking a lot about is generative AI, where that technology is developing rapidly over the next few years, and making sure that we have created the data assets to support the innovations that will occur there.

As a health system, we aren’t a software company; we don’t build foundation models or cloud platforms. But what we uniquely have is real-world, process-constrained data from care delivery and subject matter expertise to develop insights and actionable responses. This data is critical because healthcare outcomes often hinge on outliers, where synthetic datasets fall short. To that end, we’re building a number of strategic next-generation data assets, including agentic data, synthetic data, dynamic data, multiomic data, and robotic data.

How are these data assets being used? Are they primarily for internal solutions or external commercialization?

Our focus is solving Northwell-specific issues first. The goal is to improve patient care, provider quality of life, and operational efficiency. So we look to build solutions internally to address our own problems, but do so in a manner that creates the possibility of commercialization later on. To support these efforts, we work closely with Aegis Ventures, a startup studio with whom we co-create companies.

Can you share an example of a project born from this approach?

One flagship example is Ascertain, a company we co-created with Aegis Ventures, which uses agentic AI frameworks to automate backend processes. The first use case we’ve tackled is automating prior authorization processes for discharges to skilled nursing facilities or long-term care facilities. This automation has reduced processing time from something like two hours to about 10–11 minutes and allows staff to process at least four times more prior auths daily. Anecdotally, denial rates are decreasing as the system matches payer rules more accurately.

Beyond using our data for internal development, we’re also a founding member of Truveta, which pools data across health systems to support research. This effort is already yielding valuable insights in areas like COVID-19, flu trends, diabetes, and women’s health, and has transformative potential for healthcare innovation.

What advice would you give other health system leaders regarding data strategy and emerging technologies like generative AI?

Generative AI will transform healthcare much faster than most people expect—likely within the next 3-5 years. And its impact will be bigger than the iPhone and the internet in terms of scope. The pace of AI development is exponential. Technologies and solutions that seemed impossible six months ago are already solved.

Health systems need to act now. I’d recommend every health system start by, at minimum, establishing small, focused R&D teams to explore these technologies. If you wait until the technology is “ready” to start, you’ll be too far behind to catch up.

This doesn’t require massive resources. Start with a couple of internal champions who are excited about the technology. Give them 10-20 hours a week, executive support, and some funding to experiment.

Health systems need to be the leaders in how this technology is being used. We have a moral and ethical obligation to drive this conversation and ensure that these technologies are used for the benefit of patients, providers, and communities.

Are there specific steps or tools you’d recommend to help health systems keep up with this rapid pace?

Beyond creating a small team, look at tools and frameworks that make it easier to experiment. Companies like Google, Microsoft and Anthropic are releasing frameworks for agent-based systems that connect to the internet and internal systems. Even small tweaks to existing workflows can deliver value and prepare your organization for larger changes.