MedCity Influencers

The Missing Ingredients to Build Successful Enterprise AI Strategies in Healthcare

Losing sight of organizational goals and priorities and operating without a clear, supported plan can lead to a scatter-shot approach to AI investment that may duplicate investments, misallocate resources, and cause stakeholders to become cynical about AI’s value and potential.

There’s no denying it: AI in healthcare is on the rise. For example, research has shown that some AI systems can detect melanoma as well as or better than dermatologists. Recently launched AI models help providers better identify patients at-risk for serious lower GI disorders, diabetes progression, and undiagnosed diabetes. An Optum survey reports that 98% of healthcare leaders have or are planning to implement an AI strategy, which isn’t surprising, considering that AI is projected to save up to $360 billion annually in the U.S.

Yet, despite all the buzz around AI, healthcare lags behind other industries in AI-related hiring and adoption. While the FDA has sped up approvals of AI algorithms, over 75% of the cleared models relate to medical imaging. Meanwhile, 89% of healthcare leaders agree that they require special expertise to help them overcome the challenges of using AI.

The reality is that AI will continue to evolve from a novel application of technology into a must-have tool for healthcare organizations. In spite of, and even in reaction to, recent calls to slow AI’s momentum, healthcare leaders must thoughtfully and responsibly consider how AI will benefit their organizations. They must take a strategic, holistic and forward-looking approach to evaluating and implementing AI to ensure that what they deploy is aligned with organizational goals, supported and adopted by stakeholders, and, crucially, delivers benefits and impact commensurate with the level of investment.

A recipe for building successful enterprise AI strategies

A strategy provides direction, prioritizes efforts across the organization, aligns stakeholders around specific goals, and ensures those goals are backed by data and sound reasoning. As healthcare organizations—especially large academic medical centers and integrated delivery systems—undertake their own clinical AI innovation and deployment projects, a lack of strategic direction can cause resource allocation issues, stakeholder misalignment, decision-making delays, and ROI measurement challenges. A comprehensive, enterprise-level AI strategy helps the organization mitigate these risks. Despite the almost universal support for the desire to implement an enterprise AI strategy, however, many senior healthcare stakeholders are struggling with how to develop one in the first place. Here are a few things for AI strategists to think about:

  • Have clear organizational goals and priorities

AI is not a panacea: it can’t solve every problem or address every clinical or financial goal the organization may be pursuing. Effective AI strategies are aligned with, and must be subordinate to, overall organizational strategies. AI strategists must have absolute clarity about their organizations’ near and long-term priorities before they can have any chance of identifying AI investments that might help achieve them.

  • Understand the current AI landscape

From natively developed AI solutions to third-party tools to models built into applications, devices, and modalities, AI strategists need to develop a comprehensive inventory of their organizations’ existing AI investments. They should also seek to understand the status of projects that are in the ideation or development phase and have downstream potential to deliver value. 

  • Evaluate the portfolio

After inventorying their organizations’ existing AI investments, AI strategists should critically and objectively evaluate the performance of the overall AI portfolio. In this exercise, they should evaluate each asset’s level of stakeholder support and adoption, as well as its clinical and financial impact. They should then determine which assets are performing, which need additional investment, and which should be retired or shelved. For assets that are not yet “in production,” they should seek to estimate their potential value and impact. 

  • Map existing solutions to organizational goals and priorities

By connecting existing assets and investments with overall organizational goals and priorities, AI strategists will, organically, have the beginnings of an enterprise AI strategy because they will be able to show stakeholders how existing investments are already helping to achieve broader organizational goals. This mapping exercise may also reveal the need for additional investments in existing assets, whether in production or in development. 

  • Understand the broader AI landscape 

So far, AI strategists have been looking inside the organization – at the existing AI portfolio. However, in order to build out a complete AI strategy, they will also need to understand the broader AI landscape. They should be asking questions like: What are our gaps and are there existing solutions that can help us? What is their potential value and ROI? How well might they fit into our existing portfolio? 

  • Identify new AI gaps and opportunities 

With an AI asset inventory mapped to organizational goals and priorities, and an understanding of the broader AI landscape, AI strategists can begin to see opportunities in the “white space” where AI is not yet being leveraged. They should use these insights to identify targeted incremental AI investments and clearly articulate how those investments will support the organizations’ goals. 

  • Build a roadmap

With a draft of a strategy in hand, AI strategists should lay out a proposed plan – a roadmap – for investments in existing assets and new solutions: in effect, a list of projects in priority order. The roadmap should be informed by available budget, resource availability, and organizational appetite for change.

  • Get stakeholder input and support

With an enterprise AI strategy aligned with organizational goals, AI strategists should seek out input from key stakeholders, incorporate their feedback, and secure their support. Such support is critical for the ultimate success of the strategy.

Losing sight of organizational goals and priorities and operating without a clear, supported plan can lead to a scatter-shot approach to AI investment that may duplicate investments, misallocate resources, and cause stakeholders to become cynical about AI’s value and potential. By contrast, a well-devised enterprise AI strategy lays the foundation for a sustainable and impactful AI program that helps healthcare organizations better serve their patients and communities.

Sean Cassidy is CEO of Lucem Health, which he founded with Mayo Clinic in 2021 to bring pragmatic, responsible AI to the front lines of healthcare. Cassidy has 20 years of leadership experience in digital health. Prior to Lucem Health, he was the CEO of Corepoint Health and led the company through a merger with Rhapsody. He was also Louisville-based Zirmed’s (now Waystar’s) GM, Value Based Care, Charlotte-based Premier’s GM, Enterprise Analytics, and Initiate Systems’ SVP and GM of healthcare. Initiate was acquired by IBM in 2010.

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.