As AI adoption accelerates across healthcare, many organizations are struggling to move beyond pilots to measurable, enterprise-wide impact. In a recent “Becker’s Healthcare Podcast” episode sponsored by ELLKAY, Ajay Kapare, president and CEO of ELLKAY, discussed where AI is delivering real results, why many initiatives stall and how health systems can build the data infrastructure needed for long-term success.
Below are three key takeaways from the conversation.
1. AI success depends on data quality — not just algorithms
While enthusiasm for AI is widespread, execution remains inconsistent.
Mr. Kapare pointed to persistent data silos, poor data quality, inconsistent formats and inadequate governance as the primary culprits holding organizations back. “Garbage in, garbage out,” as he put it — and it’s a problem ELLKAY is designed to address at the foundation through integration, normalization and governance at scale.
Organizations seeing measurable AI outcomes are those that prioritize data integrity and accessibility before deploying advanced tools. Mr. Kapare noted that ELLKAY’s scale reflects what that looks like in practice: two-thirds of the country’s lab orders and results are supported by ELLKAY, and more than 260 million patient transactions flow through CommonWell Health Alliance, with ELLKAY serving as its technical service provider.
2. Struggling AI pilots reflect a strategy problem, not a technology problem
Despite growing investment, many AI initiatives fail to scale. According to Mr. Kapare, the issue is rarely the technology itself — it’s the absence of a disciplined, enterprise-wide data strategy.
“It’s not even the budget; it’s the discipline,” he said. “You’ve got to make sure that you are unifying the system, clean up your interfaces and get serious about governance.”
Health systems that skip these steps often find themselves unable to demonstrate ROI, particularly as budget pressures mount. Success comes from tightly aligning integration, aggregation and governance — not from layering AI tools on top of a fragile or fragmented data environment.
3. The most important investment is infrastructure for the long game
For many health system leaders, “enterprise data strategy” can feel abstract. Mr. Kapare offered a more grounded definition — one that encompasses integration, normalization, aggregation and access — and stressed the importance of thinking beyond the current wave of AI tools.
“AI is going to keep evolving. New models are going to keep coming, new vendors are going to keep coming, new partners are going to keep coming,” he added. “You will end up doing more and more. But your data foundation and data governance is paramount.”
The organizations that ultimately win, he argued, won’t be those who adopt every new technology first. They’ll be those that build a secure, compliant and scalable data ecosystem capable of sustaining innovation over a five-to-ten year horizon.
Bottom line
As AI continues to evolve, the gap between experimentation and real-world impact will come down to execution. Health systems that invest in clean, connected, well-governed data will be best positioned to unlock meaningful, scalable AI outcomes, while those that overlook the foundation risk falling short.
The post Real AI results start with a strong data foundation appeared first on Becker's Hospital Review | Healthcare News & Analysis.
Source: Read Original Article
