What healthcare CIOs believe about AI, and how they’re making it real | Viewpoint

The lightning-speed push to AI adoption has chief information officers (CIOs) juggling three enormous challenges: lightening the paperwork load for clinicians, making sense of ever-growing streams of patient data, and keeping the books balanced as margins tighten.

And while most have ambitious plans to make it all happen, ambition alone isn’t enough. Findings from a new 2025 CHIME Foundation survey reinforce this, pinpointing where AI can deliver the most value and what stands in the way.

CIOs must navigate complex challenges—data governance, cost constraints, and integration hurdles—to move AI initiatives from pilots to production. By understanding their core priorities, confronting real-world barriers, and choosing the right integration strategy, health systems can transform the AI promise into measurable, organization-wide impact.

What CIOs want from AI

CIOs view AI as the lever that shifts focus from transactional tasks to transformative care. According to the new survey, automating administrative workflows stands at the forefront, with 81 percent of CIOs citing it as their top priority—reflecting an urgent need to recover clinician hours spent on administrative tasks such as documentation, note generation, and claims processing.

Beyond paperwork relief, 70 percent of CIOs identify AI-driven clinical decision support as the logical next step, trusting algorithms to surface evidence-based treatment options, flag potential complications, and accelerate diagnostic accuracy when data volumes exceed manual capacity.

Equally pressing, 59 percent emphasize revenue cycle management as a pivotal use case—leveraging predictive billing, automated denial handling, and real-time cash-flow forecasting to secure fiscal stability in an unpredictable reimbursement landscape. A McKinsey report earlier this year found that CIOs who had adopted AI for this purpose and scaled it across their organizations were seeing 3.5 times the return on investment than those who hadn’t yet done so.

These three priorities—efficiency, insight, and financial resilience—form the foundation of the AI agenda, even as CIOs confront significant barriers to adoption that must be resolved before realizing its full potential.

What’s standing in the way of adoption

Confronting these barriers begins with addressing data privacy and security, and 20 percent of CHIME respondents cite this as a chief concern in implementation—underscoring the necessity of rigorous governance before any AI initiative can proceed.

These concerns echo an October 2024 Salesforce survey of CIOs across industries, which found that just 11 percent had fully adopted AI due to concerns over data and security. This means defining clear access controls, audit trails, and continuous monitoring to prevent unintended exposure of protected health information.

Nearly 19 percent of CHIME respondents point to the financial burden of AI deployment—from development through maintenance—as a second hurdle that too often stalls projects at the starting line.

Healthcare CIOs, however, may believe that cost is less insurmountable than their counterparts in other industries, as a January 2025 IDC report showed that 46% of their survey group saw lack of predictability in pricing to be one of their top obstacles. Regardless, nearly a fifth of healthcare CIOs still cite unanticipated integration costs, licensing fees, and the need for ongoing model support as challenges that can quickly exceed pilot budgets and erode executive confidence.

What’s more, 18 percent of CHIME respondents believe that algorithmic bias can erode clinician trust in the system. This calls for transparent model training, bias mitigation across protocols, and effective communication before full-scale or even limited-scale adoption. Without these safeguards, AI models can perpetuate bias, which CIOs recognize can undermine or permanently derail adoption.

Addressing security, cost, and safeguards in tandem lays the groundwork for choosing the right integration strategy. Once these foundations are in place, CIOs can determine how best to embed AI within their electronic health records (EHRs).

Choosing the right AI integration model

After securing governance, cost controls, and bias safeguards, CIOs must decide how to bring AI into their EHRs in a way that suits their organization’s technical maturity and operational culture.

For nearly half (48 percent) of CIOs in the CHIME survey, a native, in-platform embedding of AI in the EHR is “extremely important,” allowing insights to appear automatically in clinicians’ existing workflows without additional logins or interfaces.

Another 44 percent find it at least “somewhat important”, meaning that fewer than 1-in-10 CIOs may rather rely on third-party integration. Gartner also recently found that 47 percent of MISE CIOs alone look to vendor partnerships or managed-service models to outsource governance, bias-mitigation, and operational support to external experts, accelerating secure, compliant deployments.

Weighing these options carefully will help CIOs choose the approach that best aligns with their risk tolerance, resource availability, and long-term AI ambitions.

A practical roadmap for adoption

As organizations move from pilot projects to enterprise-scale AI, a structured approach ensures both speed and safety:

1. Launch a quick-win. Identify the most burdensome administrative workflow—such as ambient documentation or claims processing—and evaluate an AI solution at this small scale to demonstrate rapid ROI and build stakeholder buy-in before launching systemwide
2. Establish governance first. Before any data reaches AI models, define and document security protocols, privacy and bias safeguards aligned with HIPAA and enterprise compliance requirements.
3. Choose the right integration model. Native embedding for seamless clinician workflows; third-party integrations when specialized capabilities are required; managed service partnerships to augment internal expertise.
4. Measure, learn, iterate. Track significant metrics including clinician time saved and revenue cycle improvements, and use these insights to refine AI models, push successful pilots to production, and scale adoption across departments.

With a clear roadmap in place, CIOs can confidently transition from experimentation to execution—and ensure that AI delivers on its promise of efficiency, insight, and financial resilience.

There’s little doubt that CIOs believe AI offers transformational benefits across administrative efficiency, clinical insight, and financial stability. Yet realization of that promise hinges on a deliberate strategy—balancing quick-win with robust governance, choosing the right integration model, and committing to continuous learning.

CIOs have a clear charter: to adopt and scale AI that delivers the fastest return and can not only grow and flex with the needs of their organization but also make every part of it more nimble. That promise can only become reality if CIOs identify the right solution partner to turn strategy into action.

Sandra Johnson is senior vice president of client services at CliniComp.