Building AI Infrastructure in Hospitals Requires IT to Take a Lead Role

AI in healthcare: unlocking cost savings through automation and strategic involvement from IT

Thought leaders long have touted artificial intelligence (AI) and automation tools as ways for hospitals and health systems to deliver sorely needed cost savings, fill talent gaps, and improve patient outcomes and experiences. Adoption is accelerating as pressures mount: a recent BRG survey revealed that 84 percent of US providers were likely to pursue an AI-related transaction in 2025, largely driven by labor shortages and staffing costs.

Yet at most health systems, AI and automation deployment still occur in silos, which limits both the technologies’ effectiveness and an organization’s return on investment (ROI). By some accounts, 80 percent of AI initiatives fail due to misalignment between strategy, execution, and governance.

A critical factor is that the department most equipped to lead these efforts—Information Technology (IT)—often ends up on the sidelines. That leaves technology infrastructure without dedicated resources, use cases, governance structures, and enterprise-wide cooperation.

To fully realize these technologies’ potential, that must change. Here’s how.

Why Healthcare IT Departments Are Often Left Out

It seems intuitive that an organization’s IT department should play a prominent role in adopting new technology infrastructure. But that is not the case at many hospitals and health systems. Instead, tech initiatives often live within departmental silos with limited or no centralized leadership or coordination. This leads to costly redundancies, limited governance, a lack of consistent standards (e.g., for partnerships with vendors) and cybersecurity protocols, and challenges linking new technology tools to revenue improvements.

Budget is often to blame for the disconnect. If IT takes on the clinical department’s new AI implementation, and the latter department reaps the cost savings, how does the IT team secure its funding and rationalize the increase in its budget? Similarly, operations may want to retain control of an automation project to ensure it gets the credit for labor savings—which can create substantial risks if coordination with IT is limited.

This dynamic typically stems from a lingering organizational mindset that AI and automation are one-off projects rather than an enterprise-wide effort. IT teams, which have long been viewed as a reactive support function, may be especially vulnerable to such thinking. This in turn may inhibit their ability to measure ROI as it relates to discrete financial opportunities versus the number of “tickets” worked.

In the case of one large academic medical center client, the IT department’s small automation team created a standardized intake process, used templates for process mining, and was very effective at coordinating technology and building automation solutions. However, there was no prioritization or selection of use cases based on value and complexity to align with organizational objectives. Most important, the team could not track the benefit of the automations built and deployed. While the team was effective at building and deploying automations—especially given its size—it was at risk of being eliminated due to cost reductions instead of being seen as a necessary component to delivering cost savings.

ROI Is Key: Tying AI and Automation to Revenue Improvements

Organizations must see AI and automation through the lens of efficiency (e.g., staff working at the top of their skillsets and reductions in labor expenses) and efficacy (e.g., increase in net revenue or other core operational metrics). Otherwise, technology initiatives may not get the necessary buy-in from executive leadership.

That can be easier said than done. For labor in particular, many organizations still believe talent is binary—a matter of full-time employees moving in and out. They want to cut workforce but not make requisite investments in technology to help meet the mounting demand for their services.

Fortunately, there are several ways to tie AI and automation integrations to revenue improvements.

For example, revenue cycle management improvements can deliver cash savings by using technology to provide automated updates on the status of insurance claims, instead of hiring additional team members to communicate between providers and payers. Such systems can also help limit costly denials and preauthorization processes. This is especially important in light of ongoing underpayments from Medicare and Medicaid and substantial increases in care denials as the Medicare Advantage population continues to grow.

Other metrics may be tied to patient experience improvements, better clinical outcomes, or employee retention costs. At Nebraska Medicine, an AI-powered platform designed to support front-line nurses led to a nearly 50 percent reduction in first-year nurse turnover. At Ochsner Medical Center, a data-driven neural network model helps predict Clostridioides difficile infections—cutting such infections by half and saving $4 million in two years.

Establishing a Center of Excellence for Automation and AI

To involve IT, measure ROI effectively, and better manage technology capabilities, provider organizations should consider adopting a Center of Excellence (CoE) model. This entails:

• Centralized resources. Consolidate talent, methodologies, and tools to manage, prioritize, and monitor portfolios of programs and projects.
• An enterprise mindset. The CoE should serve as a hub for automation and AI expertise that spans functional areas and provides consistent means, methods, and accountable oversight for strategic projects—creating an enterprise-minded culture toward automation and AI.
• Leadership and expertise. The CoE can partner with functional areas to make the best decisions for your organization by leveraging the right solutions, from process improvements and enterprise risk management software to intelligent automation and more advanced AI/machine learning applications.

Vital to an effective automation/AI CoE is an interdepartmental governance structure that helps departments drive verifiable ROI through labor reduction and increases in accuracy and efficiency. This cross-functional framework is crucial, especially seeing as only four in ten executives and lawyers are highly confident in their organization’s ability to comply with current AI regulations and guidance.

Sound governance should encompass:

• Managing technology partners (e.g., with respect to contracts/pricing)
• Assisting departments in writing use cases and prioritization
• Assisting in building and deploying department-specific technology
• Monitoring and measuring performance and ROI
• Ensuring compliance with relevant laws and regulations
• Sunsetting and updating existing systems as necessary

IT departments may already be familiar with the CoE model—and the governance and staffing assignments that go with it—as it is similar to industry standards in how to deploy enterprise reporting and analytics.

Technology Must Be Part of Your Organizational Strategy

Hospitals and health systems cannot navigate financial and labor challenges without placing technology like automation and AI at the forefront.

Cost-cutting may be the initial driver for adopting such tools in healthcare. But long-term success hinges on the ability to track ROI tied to real-world outcomes—whether through operational efficiencies, revenue gains, or alleviating workforce pressures.

To truly unlock the value of these technologies and drive organization-wide engagement, IT must take a central leadership role. Adopting an enterprise-wide approach—supported by a CoE model—can help.

The technology is out there, and wider adoption of AI alone could save 5 to 10 percent of total annual US healthcare spend. Now is the time for providers to think big and put these new capabilities to use in a meaningful way.