The healthcare industry stands at a pivotal moment where artificial intelligence is moving beyond backโoffice automation toward genuine transformation of care delivery. Sarah London, CEO of Centene, highlighted this shift during a recent Fortune Most Powerful Women Summit conversation, noting that many current AI applications merely speed up billing or paperwork without addressing the underlying drivers of cost. She termed these uses โinflationaryโ because they often enable providers to capture more revenue in a feeโforโservice model rather than reducing expenses or improving outcomes. This observation challenges technology leaders to reconsider where AI investment yields the highest societal return, especially for organizations serving marginalized populations. The insight invites a strategic pivot: using AI not just to cut administrative waste but to redesign how and where care reaches patients, ultimately bending the cost curve while enhancing quality.
Centeneโs scale underscores why its perspective matters. Operating Medicaid plans in 29 states alongside marketplace and Medicare Advantage offerings, the $33โฏbillion enterprise serves roughly 26โฏmillion membersโa demographic London described as โamong the poorest, sickestโ in their communities. These individuals frequently encounter barriers to timely care, face higher rates of chronic disease, and rely heavily on safetyโnet programs. Yet historically, cuttingโedge health innovations have often bypassed this segment, leaving gaps that manifest as avoidable emergencyโroom visits and preventable complications. By focusing AI efforts on populations that stand to gain the most from proactive, accessible interventions, Centene aims to turn equity concerns into a source of competitive advantage and systemic improvement, demonstrating that inclusive innovation can also be financially sound.
One of Londonโs two highโimpact opportunities involves moving clinical expertise into the community through AIโaugmented support for frontline health workers. In rural or underserved locales where physicians are scarce, community health workers equipped with decisionโsupport tools can triage nonโacute symptoms, recommend appropriate selfโcare, and escalate only when necessary. Such AI systems analyze symptom patterns, vital signs, and social determinants to suggest evidenceโbased pathways, reducing unnecessary trips to emergency departments. Early pilots show that when workers receive realโtime guidance, they manage conditions like hypertension exacerbations or minor infections with confidence, freeing clinicians to focus on complex cases. This approach not only lowers acuteโcare utilization but also builds trust within neighborhoods that have historically felt alienated from the formal healthcare system.
The technical foundation for communityโbased AI relies on lightweight, interoperable platforms that can operate on smartphones or tablets, even with intermittent connectivity. Natural language processing enables workers to dictate observations in vernacular speech, while machineโlearning models trained on diverse datasets adjust recommendations to local prevalence rates. Integration with electronic health records, where available, ensures that any escalation carries relevant clinical history, and data collected during encounters feeds back into model refinement. Crucially, these tools are designed to augmentโnot replaceโhuman judgment, preserving the essential relational element of community health work. By investing in such technologies, health plans can convert fixed labor costs into scalable preventive capacity, yielding measurable reductions in avoidable admissions and associated expenses.
The second opportunity centers on empowering patients themselves to navigate early health decisions without defaulting to a clinician visit for every concern. Interactive digital toolsโranging from symptom checkers to personalized risk calculatorsโcan guide individuals through a structured assessment of their condition, propose evidenceโbased selfโmanagement steps, and clarify when professional input is warranted. When patients feel equipped to interpret early signs, they are more likely to engage in timely lifestyle adjustments or medication adherence, preventing escalation. Moreover, these platforms can tailor educational content to literacy levels, language preferences, and cultural contexts, addressing the very disparities that often limit the reach of traditional health education. The result is a more informed populace that utilizes the healthcare system judiciously, reserving costly interventions for situations where they truly add value.
Effective patientโfacing AI incorporates behavioral science principles to sustain engagement. Features such as goalโsetting dashboards, gentle reminder nudges, and gamified progress tracking motivate ongoing participation, while builtโin feedback loops allow users to report outcomes that further refine the algorithms. Privacy and security remain paramount; leading solutions employ endโtoโend encryption, transparent dataโuse policies, and optโin consent mechanisms that comply with HIPAA and emerging state regulations. Importantly, these tools are not intended to supplant professional advice but to serve as a first line of triage that directs users to the appropriate level of careโwhether that be a telehealth consult, a pharmacy consultation, or an inโperson appointment. By reducing lowโvalue visits, health systems can reallocate capacity toward highโneed cases, improving overall access and experience.
London was quick to note that pursuing these transformative uses does not render administrative automation obsolete; rather, it sharpens the financial justification for streamlining backโoffice processes. Every dollar saved on claims processing, prior authorization, or eligibility verification can be redirected toward enhanced benefits, such as expanded coverage for preventive services or reduced member costโsharing. This creates a virtuous cycle: efficiency gains free up resources that fund innovative care models, which in turn generate further savings through better health outcomes. For chief financial officers, the imperative becomes clearโevaluate AI investments not solely by the speed of transaction processing but by their impact on the total cost of care and the quality of outcomes delivered to the population served.
Distinguishing between merely efficient AI and economically transformative AI requires a nuanced measurement framework. The 2026 AI Adoption Report from Wharton HumanโAI Research and GBK Collective reveals that organizations tracking at least one growth metricโsuch as member engagement, preventiveโcare utilization, or reduction in readmission ratesโare far more likely to report strongly positive ROI than those focusing exclusively on efficiency indicators like processing time or cost per transaction. Specifically, 35โฏ% of leaders who monitor growth signals saw strong returns, compared with only 13โฏ% of those who tracked efficiency alone. This disparity underscores that AIโs value in healthcare is realized when it shifts utilization patterns toward healthier, more appropriate care settings rather than merely accelerating existing workflows.
Governance structures also play a decisive role in AI success. The same study found that 58โฏ% of enterprises that appointed a dedicated chief AI officer reported strongly positive ROI, whereas only 8โฏ% of firms lacking such leadership achieved comparable results. A chief AI officer can bridge clinical, operational, and financial perspectives, ensuring that AI initiatives align with strategic goals, adhere to regulatory standards, and are evaluated with balanced scorecards that capture both economic and health outcomes. For healthโcare organizations, creating this roleโor empowering an existing executive with clear AI oversightโcan accelerate adoption of highโimpact use cases while mitigating risks associated with algorithmic bias, data quality, and change resistance.
Beyond individual organizations, the broader market context reinforces why healthcare may serve as a litmus test for AIโs productivity promise. Major technology firms have pledged hundreds of billions of dollars to build the computational infrastructureโcloud capacity, specialized chips, and data pipelinesโunderpinning nextโgeneration AI models. The pressing question now is whether these investments translate into tangible output gains at scale. Healthcare, with its complex workflows, stringent regulations, and substantial cost pressures, offers a clear arena to assess whether AI can deliver multiโfactor productivity improvements that justify the enormous capital outlay. Success here would not only validate AIโs economic potential but also pave the way for broader diffusion into other sectors.
For healthcare leaders seeking to move beyond automation, a pragmatic roadmap begins with a clear articulation of the problem they aim to solveโwhether it is reducing avoidable emergency visits, improving chronic disease management, or enhancing member experience. Next, they should pilot AI solutions that directly involve community health workers or patients, employing rigorous mixedโmethods evaluation that captures utilization shifts, cost outcomes, and qualitative feedback. Investment in change management is essential; clinicians and frontline staff must be involved early to coโdesign workflows that feel supportive rather than disruptive. Finally, organizations should institute governance mechanismsโsuch as a crossโfunctional AI steering committee and defined metrics that balance financial, clinical, and equity dimensionsโto sustain learning and scale effective innovations.
Actionable advice for decisionโmakers includes three concrete steps. First, conduct an internal audit of existing AI projects to label each as primarily efficiencyโfocused or transformationโfocused, then reallocate resources toward the latter category. Second, establish a partnership with a communityโbased organization or a telehealth vendor to coโdevelop an AIโenabled triage tool tailored to the specific epidemiological profile of the member base, incorporating measurable goals such as a 10โฏ% reduction in lowโacuity ER visits within twelve months. Third, appoint or empower a senior leader with explicit authority over AI strategy, charging them with delivering a quarterly dashboard that tracks both cost savings and healthโoutcome indicators, ensuring that AIโs impact is visible to the board and adjustable in real time. By following these steps, healthcare entities can harness AIโs true potential to deliver better care at lower cost, fulfilling the promise that technology can serve both the bottom line and the communities they serve.