Artificial intelligence (AI) is spreading through healthcare at an unprecedented pace, faster than any other wave of health technology adoption. According to a report from Menlo Ventures, the industry has invested over $1.4 billion in AI in 2025, nearly tripling the total investment in 2024. From prior authorization to directory accuracy, AI now sits at the center of each strategic planning and board level conversations across the healthcare ecosystem.
Yet as organizations accelerate AI adoption, a fundamental reality is becoming increasingly clear: AI is only as reliable as the data beneath it. When provider data is fragmented, outdated, or inconsistent, even the most advanced models will produce flawed outputs. In a healthcare system still dominated by siloed and manual data processes, building a sound data foundation is no longer optional. It is the prerequisite for a meaningful and sustainable AI strategy.
To Properly Apply AI, We Need to Invest in our Data Infrastructure First.
Health plans and providers are increasingly applying AI to administrative workflows that were never designed for automation, such as credentialing, enrollment, provider directories, and claims processing. This is most evident in the industry’s disproportionate spend on digital technologies. Although healthcare represents roughly one fifth of the U.S. economy, it accounts for approximately 12 percent of total software spending and is adopting AI at more than twice the rate of the broader economy.
On the one hand, AI offers the potential to dramatically reduce administrative burden. With researchers finding that primary care providers (PCPs) need at least 26.7 hours per day to provide guideline-recommended primary care, AI has the potential to return valuable time to clinicians, allowing them to focus more fully on patients and on their own well-being.
On the other hand, poorly implemented AI can exacerbate existing problems. Models trained on unreliable data can generate inaccurate recommendations, introduce mismatches across systems, approve or deny claims incorrectly, and reinforce bias in ways that disproportionately affect marginalized populations. In nearly every case, the underlying issue traces back to the data layer. When provider data is incomplete, inconsistent, or outdated, AI becomes an added burden rather than a meaningful solution.
Before deploying or training any AI model, organizations must conduct a rigorous assessment of their current data environment. Today, only 43% of hospitals are routinely interoperable, resulting in fragmented data that cannot be accessed or trusted when it matters most. Payers and providers frequently maintain conflicting provider records, lack real time synchronization across systems, and struggle with mismatches in credentials, specialties, and practice locations.
AI can only scale effectively once healthcare organizations establish data infrastructure that is accurate, interoperable, and aligned in real time. Without that foundation, automation efforts will remain limited and fragile.
Provider Data Is the New Competitive Infrastructure
Advanced AI tools alone will not solve healthcare’s operational challenges. Organizations should approach AI pilots cautiously unless they already have an operational backbone capable of delivering clean, consistent, and trustworthy provider data across every workflow and system.
In this environment, provider data has become a form of competitive infrastructure. Organizations that invest in strengthening this foundation are better positioned to deploy AI responsibly and to realize long term value from automation.
The Data Foundation Problem
Across the industry, there is growing recognition that the success of AI depends on something far more foundational than algorithms or models. Trustworthy, interoperable provider data is essential. Any organization attempting to automate credentialing, enrollment, directory accuracy, or claims must first ensure that these workflows are operating from a consistent and verified source of truth.
Rather than layering new technology onto fragmented systems, the focus is on strengthening the underlying data infrastructure. This includes unifying, normalizing, and synchronizing provider information across payers, providers, and operational platforms. The objective is not to replace human expertise or decision-making, but to create an environment in which AI can be applied safely, predictably, and at scale.
By improving the data foundation, organizations gain clearer visibility into discrepancies, reduce avoidable errors, and establish a reliable source of truth that downstream automation can trust. This creates a stable platform for current AI initiatives in enrollment, directories, and credentialing, as well as for future use cases that continue to emerge.
This work is not about adding another tool to an already crowded technology stack. It is about addressing the structural issues that have constrained healthcare operations for decades.
AI Will Not Save Healthcare, Unless We Fix the Provider Data Layer First
AI adoption is accelerating, placing increasing pressure on healthcare leaders to keep pace with peers and competitors. However, integrating AI without a strong data foundation introduces significant risk. Because AI systems depend on learned information, weak or unreliable data can quickly compound errors and create problems that are costly and difficult to unwind.
AI will undoubtedly play a central role in the future of healthcare. Yet only organizations that invest in careful preparation and strategic data infrastructure will be positioned to sustain new AI models and adapt to ongoing technological change.
That preparation begins with fixing the provider data layer. It remains the essential key to unlocking AI’s true potential.
About Megan Schmidt
Megan Schmidt is based out of Grand Rapids, Michigan, and is the President and Chief Executive Officer at Madaket. Megan previously served as the Senior Vice President, Health Solutions at HealthPartners. Megan Schmidt attended Ferris State University. Megan Schmidt is a Certified Self-funding Specialist.
