In the fourth article of the series, Healthcare AI and Infrastructure Dependency, we examine why healthcare leaders must focus on enterprise infrastructure, governance, and partnerships to deploy clinical AI safely at scale.
Many organizations are rushing to adopt AI, yet too often the focus remains on the technology itself, with little attention to the integrated infrastructure required to fully enable it.
For CEOs and CIOs, the challenge is not simply choosing the right technology application; it is building the infrastructure, governance, and strategy that allow AI initiatives to launch, scale, and operate safely across the enterprise. In part four of our five-part series, Healthcare AI and Infrastructure Dependency, we outline why AI cannot succeed in healthcare without the infrastructure, systems, and partnerships that ensure safe, scalable operations.
Q: What questions should CEOs and CIOs be asking before investing in AI initiatives? What will the future look like?
There is an assumption that healthcare organizations will build agentic workflows directly on foundation models or even within the Electronic Health Record (EHR). Yet, most IT departments are already overwhelmed by day-to-day operations. Recent data suggests that healthcare IT leaders struggle with resource constraints, citing that the “technical debt” of legacy systems leaves little room for custom AI development.^1
What is actually happening inside health system procurement today looks very different from the Silicon Valley narrative. When a CIO sits down with clinical leadership, compliance teams, and the board to decide how to invest the next technology dollar, the priorities are far more complex and grounded in operational realities. While interest in AI is high, healthcare executives cite clinical safety and data privacy as their primary barriers to adoption.^2
Listening to the foundation model companies, their narrative suggests that AI is now powerful enough to automate clinical documentation, medical coding, prior authorization, and revenue cycle management. These companies argue that healthcare organizations should build directly on their Application Programming Interface (API), touting that traditional healthcare services and infrastructure companies are “middlemen” that can be bypassed.
However, the real questions to ask are: Have these companies ever successfully sold to a hospital? Have they ever navigated a health system procurement cycle? Or have they ever sat across from a Chief Medical Officer who must explain to a board why an AI system is safe to deploy in a clinical setting?
As a leading, fully integrated care enablement platform, we speak with healthcare organizations every week. They are not evaluating raw AI builds, nor are they standing up teams to integrate foundational AI into their clinical workflows from scratch. It is not even in their budget cycle. Instead, the CIO conversation is around which partner has a solution that works with their EHR, passes their security review, and can go live within a timeframe their board will accept.
Healthcare organizations are buying point solutions. When they adopt AI, they purchase a packaged product for a specific workflow (e.g., an ambient documentation tool, a coding assistant, a prior authorization automation product). They evaluate these tools the same way they evaluate any healthcare IT purchase–through compliance reviews, Electronic Medical Record (EMR) integration assessment, pilot periods, outcomes measurement, and committee approval. Research indicates that this point solution approach remains the dominant entry point for health systems.^3
"The CIO conversation is around which partner has a solution that works with their EHR, passes their security review, and can go live within a timeframe their board will accept."
This process can take months, sometimes over a year. For most healthcare organizations, the pace of AI development has far outstripped the ability to evaluate, procure, and deploy it. Hospital systems are governed by strict compliance requirements, clinical safety evaluations, board approvals, and institutional risk tolerance. Furthermore, they are often chronically understaffed in IT, deeply conservative about patient safety, and under enormous financial pressure that makes expensive build-from-scratch projects a non-starter. Many healthcare organizations remain in “education mode”, attending conferences to learn more, running small pilots, and determining where AI fits within their operations. They are nowhere near building custom AI infrastructure on foundation model APIs.
Instead, point solutions are the primary vehicle for healthcare technology adoption, addressing one workflow at a time. However, after deploying three or four disparate point solutions from different vendors, organizations can hit the integration wall. The ambient documentation tool does not communicate with the coding AI; the coding AI does not feed into prior authorization; and the revenue cycle tool is isolated from both.
This is where the IKS Health care enablement platform, powered by our proprietary domain intelligence with clinical decision support and human-in-the-loop oversight, becomes necessary.
We solve real workflow problems through a platform that integrates the services they are currently purchasing piecemeal. For every point solution a health system evaluates, IKS Health demonstrates that an integrated platform approach delivers better outcomes, lower overall total costs, and a single compliance boundary. This is how our customer base has expanded: starting with one or two services – for clinical documentation support or coding and revenue cycle management – and accelerating as AI capabilities make each service more compelling.
"IKS Health demonstrates that an integrated platform approach delivers better outcomes, lower overall total costs, and a single compliance boundary."
As an accountable partner, IKS Health meets healthcare organizations where they are, solving real workflow problems with a platform that integrates the services they are buying. We ensure every point solution, from ambient documentation to revenue cycle tools, contributes to a unified, connected care journey.
Q: What will responsible AI look like years from now?
Foundation models will continue to evolve. At IKS Health, we welcome this progress because advanced models enhance our platform, driving higher-performing coding models through our integration and oversight infrastructure. This synergy produces results that are more accurate, faster, and less expensive.
Could AI regulation become more streamlined, potentially reducing the need for a deeper compliance layer? Or will the proposed Health Data, Technology, and Interoperability (HTI-5) rule signal a pivot for AI and healthcare tech companies? The jury is still out. Current regulatory proposals seek to utilize the HTI-5 framework to scale back transparency requirements and prioritize Fast Healthcare Interoperability Resources (FHIR) based interoperability, potentially easing the burden for integrated platforms.^4
If regulations lean toward loosening controls, it could unlock an opportunity for AI companies by providing frictionless API access to EHRs, enabling autonomous agents to schedule care, draft clinical notes, and optimize workflows at scale. This would accelerate innovation, lower barriers to entry, and shift value toward data orchestration and intelligent automation layers. If controls tighten instead, incumbents may retain their moats, slowing agent adoption but preserving stability, auditability, and clearer accountability. In either scenario, CEOs and CIOs will need to navigate a shifting liability landscape by focusing less on any single AI model and more on the infrastructure that allows many AI initiatives to launch safely and repeatedly.
Could healthcare organizations build their own platforms internally? The financial case for building erodes over time as the established platform compounds its data advantage and compliance requirements grow more complex. The build option may have seemed plausible when healthcare AI was limited to one or two workflows, but the scale has shifted with emerging services and the necessity of continuous compliance maintenance. The platform economics are becoming overwhelming. Current industry analysis suggests “buy” is winning, as organizations now prefer third-party vendors over in-house builds to mitigate the high costs of ongoing maintenance and security.^5
The trajectory of AI and foundation model improvement directly benefits IKS Health. Because AI can never be an accountable partner on its own, IKS Health provides the essential layer of assurance through our integrated care enablement platform, ensuring we are accountable for each AI application.
The future is a mature ecosystem where many AI systems operate simultaneously and continuously. As the AI hub, IKS Health will continue to lead by building systems capable of supporting hundreds of AI-driven processes operating across the organization, ensuring that clinical quality remains the priority.
Sources:
^2. KPMG Intelligent Healthcare 2025: A blueprint for creating value through AI-driven transformation (KPMG)
^3. Healthcare AI Update 2025: What use cases are adopted the most (KLAS Research)
^4. ASTP/ONC’s year-end moves mark a strategic pivot in federal health IT policy (Holland & Knight)
^5. The pros and cons of buy versus build (Healthcare IT News)