Beyond Automation: Why Healthcare’s AI Moment Demands a Higher Ambition
-
September 15, 2026
-
Health systems face pressures from workforce shortages, margin compression and an increasingly complex regulatory environment. But as systems race to deploy AI to address these challenges, they must avoid using new technology to accelerate processes that no longer create value.
When AI is bolted onto inadequate or outdated processes, risks increase and expected gains quickly evaporate. To realize meaningful value from AI, health system leaders must move beyond asking, “How do we do this faster?” and instead ask, “Should we do this at all?” They must leverage AI to challenge assumptions, eliminate low-value activities and redesign how work gets done.
Governance must evolve alongside AI, as it transforms from a productivity tool into a decision-enablement platform. Though cost and productivity gains matter, strong AI portfolios balance economics with clinical outcomes, workforce sustainability and enterprise risk.
Health systems that use AI as a catalyst for process reinvention will gain significant advantages in building a foundation for cost reduction, operational efficiency and resilience.
AI as a Redesign Catalyst
Healthcare systems continue to view AI largely as an accelerator to do work faster and with fewer people. While true for certain use cases like prior authorization processing, clinical documentation and revenue cycle coding, this view ignores the possibility that the underlying processes may be poorly designed.
Beyond cost reduction, AI’s greater potential is to challenge legacy assumptions and create entirely new operating models that drive business value. The distinction between automation and reinvention is already evident in some of healthcare’s most prominent AI use cases.
Clinical documentation, for instance, has long been a cumbersome and distracting process for physicians. The traditional model requires physicians to document clinical reasoning into structured notes, a process frequently cited as a key contributor to the practitioner burnout epidemic.
New AI documentation tools not only transcribe faster but also enable better patient interaction and provide other positive downstream impacts on care continuity. By integrating documentation into the clinical encounter itself, these tools reduce administrative burden and allow physicians to better focus on patient care.
The opportunity extends well beyond generative AI and administrative workflows. Predictive models are changing when and how clinicians intervene. By identifying high-risk patients earlier, physicians can help prevent patient deterioration and crises.
A large nonprofit health system is using predictive models to identify patients at risk for sepsis and heart failure, producing measurable reductions in mortality and readmission rates that process automation alone could not have achieved. Without these models, proactive intervention and outreach before a patient’s condition worsened would have been difficult.
But as models evolve from automating tasks to making recommendations, systems must balance innovation and governance to manage risk and maintain trust.
The Cost-Takeout Trap
The AI benefits with the greatest impact are often the hardest to quantify and may not affect the bottom line. But more accurate diagnoses, greater consistency in care delivery, stronger patient experiences and a sustainable clinical workforce can generate compounding benefits for both patients and healthcare organizations.
In addition, focusing solely on short-term objectives can create unintended behaviors and consequences. A health system that evaluates its AI investments solely based on labor cost reduction, for instance, will naturally prioritize applications that displace headcount. While this may be helpful to the chief financial officer and the health system’s net income, the singular focus ignores investments that might improve and augment clinical judgment and ultimately patient outcomes.
The evidence on AI-assisted diagnostics illustrates the gap. Google’s DeepMind developed an AI system for detecting more than 50 eye diseases from optical coherence tomography scans, matching or exceeding1 the diagnostic accuracy of expert ophthalmologists. The value of that capability is not primarily that it reduces the cost of reading a scan. It makes expert-level diagnostic accuracy available at a scale and speed the existing specialist workforce cannot match — extending access to populations and geographies that would otherwise go unserved. Framing that as a cost story misses the point entirely.
Process Evaluation as a Precursor to AI Consideration
Healthcare providers have invested significantly in AI and predictive models to identify patients at increased risk of adverse clinical outcomes, with mixed results. In some cases, AI deployments have exposed or amplified gaps in underlying clinical processes. To be successful, AI deployment should include robust processes for validation, escalation and clinical oversight, in addition to monitoring mechanisms.
Given the inherent risks of AI automation, healthcare organizations need to intentionally and deliberately evaluate risks, risk tolerance and controls before adopting an AI operational strategy. Internal conversations should start by understanding and evaluating the workflows or operational processes in place and then determining whether and how AI can complement or supplement them.
Evaluating the suitability of AI for a particular operational process requires a cross-functional risk assessment. The assessment should examine workflows, decision criteria, accountability, risk tolerance and the control environment, and address several key questions:
- Does the operational process in question need to exist at all, or does technology advancement enable the elimination of the existing process or consolidation of processes to accomplish the desired objective differently in the future?
- Are there opportunities to simplify existing operational workflows to remove unnecessary complexity and risk before AI deployment?
- Does the company have a well-reasoned, shared risk tolerance specific to the operational risks in that area, and is that tolerance widely understood?
- What controls exist in the current process? Are those controls adequate and effective in monitoring the risk, and will implementation of AI technologies render those controls ineffective or inadequate?
Governance processes designed to consider AI risk must give specific attention to these considerations before moving on to discussions of how and when AI deployment will take place, or what type of AI solution is appropriate for a given need. Organizations should make these conversations a mandatory part of early AI decision-making, especially in clinical operations, to enhance risk management and support successful strategic outcomes.
A Higher Standard for Healthcare AI
While it is appropriate and necessary to embrace AI adoption in healthcare, healthcare organizations and their advisors have a responsibility to hold AI adoption to a higher standard. That means asking, before any implementation, whether the process being automated is the right process and whether the risks of AI adoption are well-understood and effectively controlled. It means measuring AI value in terms that include clinical outcomes, workforce experience, and access — not only cost per transaction. And it means recognizing that the most valuable thing AI can do in healthcare is not to make 2010’s operating model run faster, but to make possible care delivery models that 2010’s technology could not have supported at all.
The views expressed herein are those of the author(s) and not necessarily the views of FTI Consulting, Inc., its management, its subsidiaries, its affiliates or its other professionals. FTI Consulting, Inc., including its subsidiaries and affiliates, is a consulting firm and is not a certified public accounting firm or a law firm. FTI Consulting is an independent global business advisory firm dedicated to helping organizations manage change, mitigate risk and resolve disputes: financial, legal, operational, political & regulatory, reputational and transactional. FTI Consulting professionals, located in all major business centers throughout the world, work closely with clients to anticipate, illuminate and overcome complex business challenges and opportunities. ©2026 FTI Consulting, Inc. All rights reserved. fticonsulting.com
Footnotes:
1: De Fauw, Jeffrey, et al. , “Clinically applicable deep learning for diagnosis and referral in retinal disease,” Nature Medicine (Aug. 13, 2018) , 1342–1350.
Related Insights
Related Information
Published
September 15, 2026
Key Contacts
Senior Managing Director
Senior Managing Director
Senior Managing Director