Achieving sustainable value with healthcare AI

As AI continues to reshape the operations of hospitals, health systems, and payers, leaders are increasingly seeking ways to determine the viability and premium returns of AI-driven initiatives. More often than not, these organizations rely on benchmarking methods—comparing performance metrics to industry standards or peer institutions—to assess new technologies, with a focus on elevating performance and achieving operational results. 

Executing against traditional operating measures and data-driven benchmarks is fundamentally time-based and is increasingly limiting in this new era of technology. To achieve more sustainable performance and compounding returns, leaders should evaluate AI use cases against the viability and feasibility of the experiences they hope to create. 

The multi-dimensional value of target state experiences in AI investment decisions 

A majority of healthcare organizations are using some form of AI across their operations in their daily operations, mostly in back-office and administration functional use cases. Seventy-eight percent of organizations have already implemented or are currently implementing AI into workflows, according to a Guidehouse-HIMSS survey. Whether the investment is for automation or augmentation, benchmarking for the purpose of qualifying investment decisions is typically based on quantitative metrics associated with performance controls across cost, efficiency, quality, and safety. While valuable for identifying gaps and setting improvement goals, benchmarking has notable limitations when applied to assessing AI use cases: 

• Time-constrained: Benchmarking references historical patterns and compares existing performance to present-day industry standards versus an organization’s target-state strategic identity, thereby limiting leaders to incremental progress instead of transformative change. 
• Outcome bias: Traditional industry benchmarks are not easily customized to the unique needs, populations, or strategic priorities of a specific healthcare provider or payer. This makes it more difficult to holistically evaluate AI’s impact on employee, physician, patient, and member outcomes. 
• Masked value: Organizations that set goals solely on historical performance indicators to meet industry competitive performance risk missing innovative and industrydefining solutions that can reset benchmarks and position them as trailblazers across care delivery, operational excellence, and growth frontiers.  

Healthcare organizations that shift to an experience-centric approach are establishing the foundation for the long-term, compounding returns of AI investments and defining the optimal future scenarios for all stakeholders involved: patients, clinicians, administrators, and the broader community. For instance, an AI tool that detects lung nodules in radiology images would typically be evaluated on how precisely it identifies abnormalities and how quickly it processes scans. While these metrics validate performance, they fall short of capturing real-world impact.  

Tomorrow’s measurement framework must go further, focusing on outcomes that matter: Did earlier detection improve survival rates? Did the tool reduce downstream costs by avoiding late-stage treatments? Was clinician confidence enhanced, and did it alleviate burnout? Did the application advance equity by improving care for diverse populations? By shifting from algorithm-centric KPIs to patient-centric and system-level metrics, healthcare can validate that AI delivers value that is truly fit for purpose. As a forcing function, this will drive the evolution and maturation of the metrics required to effectively capture performance. 

This vision-driven approach starts by determining the experience of AI, versus only planning for the quantitative impacts of implementing AI as a performance accelerator, with a focus on: 

• Stakeholder centricity: By indexing on desired experience for stakeholders across their ecosystem, healthcare organizations can identify AI use cases that directly address pain points, improve satisfaction, and drive better outcomes. 
• Strategic orientation: Target-state design methods align AI initiatives with long-term strategic priorities, rather than chasing industry benchmarks that are based on the present.
• Reinvention trajectory: Designing for experience in healthcare encourages unconventional thinking, opening the door to radically new models of care, operations, and workforce strategy. 

A strategic orientation to yield better returns 

When determining what AI investments to make, healthcare organizations should start with the target state experience to yield higher returns—both financial and non-financial—compared to standard benchmarking for several reasons: 

• Distinct value creation: AI solutions are tailored to the organization’s specific context, maximizing the likelihood of AI adoption and stakeholder impact 
• Elevated buy-in: Stakeholders are more engaged when changes reflect their needs and aspirations, reducing resistance and accelerating implementation 
• Sustainable performance measures: Instead of common benchmarks, leaders define success in terms of the improved experiences and outcomes they seek in a broader digital ecosystem 
• Fit-for-purpose: Target state experiences help executives anticipate and adapt to future trends, regulatory changes, and patient expectations, reducing the risk of becoming obsolete in a rapidly changing competitive healthcare landscape 

When compared to AI investment approaches that evaluate AI use cases against benchmarks and either new or existing technology applications, organizations that embrace the value of human and technology intervention are better positioned to be “Market Setters” (see Figure 1). 

 

Figure 1: AI Investment Identity Spectrum 

Activating experience-based use cases for AI  

To adopt a target-state experience approach, healthcare organizations should: 

• Start with the value of experience: Healthcare organizations should define what the experiential impact of AI can do for patients, clinicians, and administrators across the ecosystem, in order to inform the use cases and applications that will position them for surpassing standard benchmarks and leapfrogging the competition. 
• Balance performance needs within a sustainable model: The healthcare organizations that will exceed performance standards while elevating experience over time are those that invest in AI use cases in the context of the future operating model and infrastructure, versus only in functional gaps that enhance back office and front office performance standards today. 
• Customer “market-setting” metrics: Organizations should customize operating and performance metrics that reflect progress toward the target-state experiences, versus only near-term gains. Rather than defaulting to metrics like cost, quality, or time-saved, this will require a more intentional evaluation of the specific expected outcomes for each use case.  
• Elevate the human dynamic: While AI use cases are often compelling, they can place disproportionate emphasis on automating human tasks, reinforcing stigma around AI and diminishing the perceived value of workforce roles. Instead, healthcare leaders should focus on amplifying the human experience in ways that guide AI investment decisions and elevate the value of human judgment and intervention. 

The outlook 

While benchmarking remains a useful tool for continuous improvement, traditional metrics are insufficient for guiding the transformative potential of AI in healthcare. By starting with target-state experiences, healthcare organizations can qualify that their investments in AI are strategic, value-accretive, and truly aligned with the needs of those they serve. This approach not only delivers greater returns but also fosters a culture of innovation and excellence that will define the future of healthcare.