Data & TechnologyThe freewheeling days of healthcare AI spending are over, but not because AI failed to deliver. Health systems remain bullish on AI use cases and their ROI potential. What’s changing in 2026, however, is how those investments are evaluated.
As providers face a multitude of macroeconomic concerns, IT budgets are coming under scrutiny, and money for AI governance is at risk. Only 26% of hospitals say they plan to raise AI governance and safety budgets by 2% or more in 2026, according to a Black Book Research survey. Another 18% plan no increase.
These budgetary pressures mean most new AI ideas won’t get approved unless teams have a cost-effective governance strategy. In this article, we’ll explore two forces driving fiscal concerns, unpack where governance shortfalls exist, and reveal steps IT leaders can use to advance AI in their organizations.
Why IT budgets are under the microscope
Health systems continued their post-pandemic recovery last year, with health system margins growing for three consecutive months from July through October 2025. In 2026, however, health systems, hospitals, and providers face a new wave of economic and regulatory challenges that are set to squeeze IT budgets.
Changes due to the OBBBA
The One Big Beautiful Bill Act (OBBBA), also known as H.R. 1, which was signed into law last July 4, brings sweeping changes that will impact health systems in 2026. On the positive side, the OBBBA will increase funding to rural health systems. On the negative side, the OBBBA is projected to have serious healthcare-related economic consequences, including increased uncompensated care and bad debt, reduced Medicare rates, and less funding from government programs.
Macroeconomic concerns
Healthcare workforce shortages and inflationary increases in wages and supplies place additional pressures on IT budgets. These concerns aren’t stopping AI investment, however. AI is a top priority for 45% of healthcare leaders in 2026, according to a Deloitte survey of 59 health system executives. Respondents say those investments will cover three areas:
- Back-office functions, including clinical documentation, revenue cycle management, and workflow management
- Business functions, including patient experience, care management, marketing, engagement, and regulatory compliance
- Clinical and care enablement functions, including diagnostic imaging, pathology, and clinical decision support
The tension between lower IT budgets and greater expectations for AI will require IT teams to prove that their AI investments are low-risk and high-ROI. That proof depends on governance, but recent data shows that most health systems are still missing the governance foundations to consistently scale AI. Let’s see why.
The current state of AI governance in health systems
While awareness of AI governance is growing, execution lags behind ambition. Last year, 70% of hospital CFOs said their organizations had at least some governance structure, up from 40% in 2024. Even so, results from the Black Book Research survey show that most organizations still face significant gaps in their ability to scale AI and achieve ROI.
Inconsistent governance policies
While awareness of AI governance is growing, execution remains uneven. Only 29% of hospitals have fully implemented and enforced policies covering areas such as AI model inventory, lineage, and formal sign-offs. Nearly half (48%) are still drafting such policies. Without consistent standards, AI initiatives are evaluated piecemeal, slowing approvals and increasing the risks of stalled pilots.
Unclear ownership across teams
Governance efforts often break down because responsibility is fragmented. One-third of hospital leaders say unclear ownership among IT, quality and safety, and compliance teams directly impedes their AI governance efforts. When no single group is accountable, finger-pointing begins, and progress stops.
Limited auditability and transparency
To build trust in AI, leaders need full visibility into model decision-making, but too often, that doesn’t happen. Just 22% of hospital leaders are confident they could produce a 30-day audit trail for regulators or payers. Among smaller hospitals, that figure drops to just 15%. Vendor practices compound the problem. Forty-one percent of executives cite limited documentation — such as model cards and drift reports — as a top barrier to audit readiness.
Lack of formal AI governance councils
Many hospitals still lack a centralized forum for AI oversight, keeping them in the slow lane when it comes to achieving ROI. Health systems with a formal AI governance council achieve ROI twice as fast as those without one.
These four challenges cause real-world problems. Another Black Book Research report shows that 70% of healthcare leaders report at least one failed AI pilot due to governance-related concerns. Additionally, 80% of executives say vendor AI claims are difficult to verify without formal governance.
Steps IT teams can take to improve governance and prove ROI
These issues may seem difficult to address in an environment of limited AI governance funding. Fortunately, there are steps IT leaders can take to reduce risk and increase ROI while keeping governance costs in check.
1. Choose the right workflows
When IT teams implement AI into well-defined administrative workflows, such as intake or referral, governance is easier because there are fewer unknowns. By contrast, applying AI to loosely defined or judgment-heavy workflows like predicting clinical outcomes or predicting treatment pathways introduces risk.
One prime example of a well-defined, AI-ready workflow is the referral process, a traditionally unwieldy, paper-heavy, and arduous task. The scope of work is clear, the inputs are known, and success can be measured against benchmarks such as faster turnaround time or reduced referral leakage.
2. Invest in clear governance structures
AI governance councils and other oversight bodies save IT teams from duplicative work that makes it difficult to scale AI cost-effectively. Establishing a single decision-making entity means IT doesn’t have to re-explain each project, legal and compliance don’t have to perform redundant reviews, and security teams can apply pre-approved protocols instead of redefining safeguards for every new deployment.
3. Choose lower-risk implementations
Not all AI implementations require the same level of oversight. Clinical AI, which directly impacts diagnosis and treatment, is high-risk, consuming multiple governance resources. In contrast, operational AI tools don’t make clinical decisions, bringing lower compliance and validation costs. Operational AI tools are also easier to justify, audit, and turn off if needed.
4. Make change management simpler
Adoption is a prerequisite to achieving ROI. That’s why IT leaders should choose AI tools that create less friction for users. If an AI platform has a separate UI and requires a new login or password, it’s less likely to achieve mass adoption by employees and end-users than one that’s seamlessly integrated into a health system’s existing solutions.
Intelligent Document Processing: A Governance-Friendly Approach
While some AI implementations can be complex to govern, healthcare-specialized intelligent document processing (IDP) solutions — leveraging technologies such as large language models (LLMs), natural language processing (NLP), and optical character recognition (OCR) — are designed with governance in mind. These tools transform unstructured patient information into structured documents with clear audit trails.
As an example, modern intelligent data extraction tools can:
- Extract patient name, birthdate, ICD-10 codes, and other critical information from paper faxes, PDFs, scanned documents, and even handwritten doctor’s notes
- Populate them into a structured format that meets HL7/FHIR specifications for secure healthcare data exchange
- Route structured documents into a system of record (EHR, RIS) seamlessly without requiring management in another platform
With extractive AI working behind the scenes, health systems can transform cumbersome processes, such as prior authorization. Staff spends less time on manual data entry. Document processing times decrease. Prior authorization paperwork is completed in accordance with CMS guidelines. And more referrals remain within the health system.
How eFax® Clarity simplifies AI governance
eFax® Clarity is a high-impact, budget-friendly intelligent document processing tool that simplifies backend work. Built to improve operational workflows, eFax Clarity lets organizations take a low-risk approach to AI adoption without breaking their limited governance budgets by providing:
- High accuracy. Because eFax Clarity pulls data from existing sources, it produces predictable, deterministic outputs tied to source documents.
- Traceability. When integrated with eFax, eFax Clarity provides centralized dashboards. Administrators get full visibility into inbound and outbound activity. They can also manage users and track usage. eFax Clarity also generates a confidence score that health systems can use to monitor system performance.
- A defensible audit trail. Information is easy to trace, with auditability that confirms the source and destination of all extracted information.
- Exception handling. When extraction errors occur — whether due to illegible handwriting, incorrect form completion, or a scheduled manual audit — the Exception Management Portal offers a streamlined solution. Through this intuitive web interface, your staff can review and correct data before it ever reaches the EHR. By establishing custom rules based on confidence scores, you can automate which documents require a human eye, ensuring that exceptions are resolved and submitted with just a few clicks.
- Multi-channel document support. eFax Clarity can process documents from multiple sources, not just faxes, but also emails, scanned documents, and direct uploads, providing flexibility in how your organization receives and processes patient information.
Because it focuses on structured data extraction with clear audit trails, eFax Clarity introduces lower governance overhead. It’s also easy to integrate, supported by expert teams that can adapt it to any health system’s tech stack. And because there’s no separate UI or user login, employees can leverage the benefit of intelligent document processing within the workflows they already trust.
Invest in AI that holds up under scrutiny
AI budgets may be going under the magnifying glass in 2026, but that doesn’t mean health systems can’t find ways to strengthen governance and improve operations. Investing in low-risk tools designed to work with existing systems will deliver measurable, consistent results.
Ready to see how AI-powered intelligent data processing can make intake, referral, and referral management processes easier?
Learn more about eFax Clarity.
