Artificial intelligence is rapidly reshaping how healthcare works. It promises to streamline documentation, automate scheduling, reduce administrative burden and help patients navigate systems that too often feel fragmented and frustrating. For clinicians and staff long stuck performing tasks that are repetitive, process-driven and poorly suited to human judgment, these changes are genuinely exciting.

But stepping back from that excitement raises a harder question: Will faster processes and smoother navigation meaningfully change what drives the rising costs of chronic disease, behavioral health needs and member disengagement?

AI creates capacity. The strategic choice is whether that capacity will be used simply to optimize transactions, or to invest in what actually improves health.

If the time and resources freed by automation are reinvested in sustained human relationships, AI could become more than a cost-control tool. It could enable health plans and providers to build roles centered on empathy, continuity and trust — factors that often determine whether someone manages a chronic condition, follows through on care or feels supported in moments of vulnerability.

The Relationship Is The Intervention

In one randomized trial conducted during the pandemic, older adults receiving regular empathetic phone calls from trained laypeople reported significantly lower loneliness, depression and anxiety than those receiving usual support.

A subsequent randomized trial among low-income patients with diabetes found that similar empathy-focused outreach was associated with meaningful improvements in glycemic control compared with typical care.

A growing body of research — including randomized trials of peer support, peer health coaching and community health worker interventions — shows that social connection, lay engagement and structured support can improve mental health, engagement and chronic disease outcomes across multiple populations.

At first glance, these findings may seem intuitive: people do better when they feel supported. But the science shows something more significant. When applied with purpose, relationships can have a measurable impact on specific health conditions that the healthcare system otherwise struggles to influence.

For health plans and providers, these improvements translate to better management of chronic disease and earlier intervention when necessary, which in turn lowers healthcare utilization and supports a stronger financial performance for health systems and plans.

Designing for Human Connection at Scale

If the relationship is the intervention, the next question is how we make that intervention available to many more people without worsening workforce burnout. AI gives healthcare providers room to maneuver by taking on the repetitive, process-driven work humans currently do by hand. But the way organizations choose to use that room will determine whether AI leaves healthcare more humane, or simply more efficient at doing what it already does.

There are several practical ways to start. Medicare Advantage plans, Medicaid managed care organizations and provider groups operating under shared savings contracts are uniquely positioned to deploy this model. AI can identify members at rising risk, summarize longitudinal data, and prompt outreach. The intervention itself — a structured, ongoing human relationship — can be delivered by trained lay guides or community health workers, supported by technology rather than replaced by it.

This approach invites a rethinking of where clinical expertise is truly required. Research has shown that with minimal training and support, laypeople can reliably deliver empathic contact that improves mental and physical health. So rather than directing every AI investment toward tools for licensed clinicians, systems could pair automation with recruitment and training of guides or peers — people who may share language, culture or lived experience with those they support, or simply be talented in building trusted relationships. AI can help manage caseloads and flag risk, but the core of the health-influencing work is a human relationship.

Choosing Wisely in the Age of AI

In the era of AI, we must carefully consider where we choose to scale with AI’s help. A system that uses automation primarily to cut staff or speed up billing will likely see gains on a spreadsheet but little change in people’s lives. A system that uses those same efficiencies to create space for sustained, empathic contact can aim for better health outcomes.

For example, a health plan might decide that every member with a new chronic disease diagnosis should receive a series of non-clinical check-in calls for a few months, with AI helping to capture relevant insights and report them back to the care team and plan. A primary care practice might allocate part of the time freed up by AI-enabled documentation to extend visits for patients navigating serious illness, or to follow-ups from a team member whose job is to ask, “How are you really doing?”

Early studies demonstrated what can happen when human connection is treated as a core part of care. AI now offers a way to build on that insight at scale by making it easier to offer relationships consistently to the people who need them most.

The question facing healthcare leaders now is whether risk-bearing organizations will use AI’s efficiencies to build the human infrastructure that member health challenges demand. If the relationship is the intervention, then payment models, workforce design and AI investment strategies will need to reflect that reality.