Forum Insider - June 2026 - Forum Season Report Out

Consumer AI tools are raising questions about how health systems acquire patients and build long-term relationships, with some early signals suggesting patient behavior may already be shifting — particularly among younger populations.

One large integrated system observed a notable decline in under-50 patient engagement over six months, which leadership attributed to consumer AI tools — though peers debated whether this trend is motivated by convenience, cost, or a genuine shift in trust.

The shift raises existential questions about long-term capital: systems that reoriented their footprint around virtual care are asking whether they're falling behind, and what a “health system of the future” would look like built from scratch.

Health systems are accelerating ASC investment, but the right approach to equity, branding, geography, and governance varies significantly by market context.

Physician equity is increasingly seen as a key alignment lever. One system attributes its large ASC portfolio directly to physician equity participation; another changed its policy after identifying the restriction as its biggest growth bottleneck.

In-market ASC growth can carry cannibalization risk — one system noted volume displacement that “has not gone well.” CFOs are wary, but many CSOs see the outpatient shift happening regardless.

Despite broad strategic alignment on ambulatory expansion, ambulatory investments frequently compete within the same general capital pool as hospital infrastructure — a governance dynamic several leaders flagged as a growing constraint.

The bottleneck can be governance, not vision: several systems reported physician buy-in for ambulatory growth but no dedicated capital pathway, forcing ambulatory to justify itself against hospital projects that appear more immediate.

Construction costs ranging from $1,000 to $2,300 per square foot, with land exceeding $1 million per acre in high-growth corridors, complicate the ROI of footprint growth.

With major programs like 340B under scrutiny and traditional advocacy drawing diminishing returns, strategy leaders are reconsidering how to shift from a largely defensive posture to proactive engagement with policymakers.

The traditional approach of opposing every unfavorable reform proposal may need to give way to bringing constructive alternatives that align with the current administration's priorities.

A fragmented industry — for-profits, academic centers, rural hospitals, nonprofits — operates under different economics, making a unified message difficult to sustain.

Some systems are beginning to treat service lines as distinct products with tailored governance, marketing, and workflow design — an approach generating early traction while exposing how much existing structures were built around physician convenience rather than consumer needs.

End-to-end workflow redesign across virtual, ambulatory, and specialty settings is being pursued at some systems, though it requires sustained coordination most structures don't yet support.

Physical therapy illustrates how strategic framing shifts viability: one system treats it as a declining cost center, another as a consumer acquisition channel feeding higher-value services.

Several systems are recovering capacity through virtual care, scheduling analytics, and workflow redesign rather than new construction — though maturity levels vary.

One system's care-at-home model freed 19 beds that backfilled immediately with higher-acuity patients, while serving 4,500 patients annually at a steady-state daily census of 100. Leaders framed it as cost avoidance rather than a revenue line.

One system deploys virtual behavioral health visits in K-12 schools, college campuses, and field settings, using the privacy of a remote visit to reduce stigma and improve access.

The AI conversation sounded more like operations than experimentation. Systems further along run formal governance and treat clinician feedback as a veto, though most remain early.

Time savings are the most defensible early metric — one system saved 1.6 million nursing hours — though leaders cautioned that vendor savings claims often contain “fuzzy math.”

Data actionability, not availability, is the constraint. One leader described being “data rich but information poor,” noting that tools surfacing existing data are proving more valuable than new data collection.

The drug discount program is under pressure from an industry-funded messaging campaign, a proposed shift to a rebate model, and accelerating manufacturer restrictions on contract pharmacies.

The proposed rebate model could cost hospitals over $1 billion by converting upfront discounts to post-pay reconciliation — effectively interest-free loans to manufacturers. A judge blocked the first pilot; a relaunch is expected in 2027.

Site visits have been the most effective advocacy. Several leaders described legislators tuning out during policy briefings but responding to direct visits to funded programs.

An investment framework separating health systems into mission, regulated hospital operations, and growth-focused business lines gave the room vocabulary for tensions many already felt.

Running all three under a single operating model masks the economics of each, several leaders agreed — though how to restructure remains contested.

Leaders discussed a potentially closing first-mover window on converting interchangeable outpatient sites to non-facility pricing before competitors strike deals.

The workforce conversation moved from filling vacancies to designing roles people stay in, with physician partnership models and AI as central enablers.

One CRNA partnership dropped attrition from 25% to 6% through job design — schedule predictability, top-of-license practice, professional respect — not compensation.

The anesthesia labor crisis is a 10-plus-year structural problem, with several leaders describing the cycle of rising subsidies and opaque cost drivers as unsustainable.

Leading systems are assessing what the revenue cycle organization needs to look like when a meaningful share of today's transactional work runs autonomously.

Vertical siloes — within RCM and with clinical leaders — need to dissolve to capture AI ROI. Many systems are actively restructuring in favor of a more unified, horizontal revenue team and are reporting meaningful operational improvements as a result.

Leading systems are requiring teams to fund new tools by removing cost from existing budgets rather than seeking incremental capital — putting budget at stake to sharpen tool selection.

The limiting factor in AI adoption remains the workforce, not necessarily the technology. The shift from task execution to analytical oversight is requiring more change management investment than most organizations anticipated.

AI may prove inflationary rather than deflationary if workforce transformation isn't treated as a first-order priority. Role redesign is currently outpacing workforce readiness.

Staff anxiety about displacement is slowing adoption even where messaging has been deliberate. The systems making the most progress have reframed the conversation as job reshaping, not job loss — and are having that conversation continuously, not once.

Leading systems are embedding revenue cycle into capital planning, service line design, and partnership decisions from the outset — resolving billing model and managed care assumptions before go-live.

The “copy/paste” versus “file new” framework is critical: most revenue cycle go-live failures stem from treating a genuinely novel capability (new payer products, risk-bearing models) as if it were a routine replication.

MA exits, ICHRA proliferation, and direct-to-employer arrangements all carry revenue cycle consequences invisible in a standard capital submission. Benefit design is now a core revenue cycle competency.

No broadly investable solution has emerged, and the structural conditions are moving against both providers and payers. Systems are nearing a choice: pursue genuine payer collaboration, or build the bypass routes that reduce dependence on the relationship altogether.

Litigation resolves specific claims but does not change payer behavior. Durable change requires contract language reform.

Consumer AI is an emerging competitive threat. Several systems reported AI agents calling on behalf of patients to manage accounts — with no operating protocols in place for it.

MA exits, the pivot toward employer-sponsored and direct-to-employer arrangements, and the trajectory toward partner-driven revenue are collectively reshaping the base that revenue cycle was optimized to manage.

The payer mix shift is a design problem, not just a contracting one — getting ahead of it requires revenue cycle involvement in purchaser strategy.

In competitive markets, a poor patient financial experience is enough to lose a patient permanently. Systems are beginning to treat revenue cycle’s consumer-facing functions like estimates, self-service, and digital payment as part of the growth strategy, not just the back office.

One system collapsed eight separate CV leadership structures into a single P&L owner, reporting faster decisions and a clearer investment narrative for the C-suite — though the transition required senior buy-in and over a year of cultural work.

An equity-free, salaried physician model can remove site-of-service incentive distortion: one system's CV ASC generates strong margin with patient placement driven by clinical appropriateness rather than physician economics.

The systems reporting the most integration progress invested heavily in relationships before pursuing standardization — and those that skipped that step described resistance that stalled progress for years.

One academic system spent its first year on cross-campus trust-building with a minimal budget, deferring standardization until leaders at each site felt valued.

Local variation can be a feature, not a defect. Attempts to homogenize sites that evolved for different markets risk destroying what makes them functional.

CV leaders have moved past treating AI as a future consideration, but the binding constraint is the workflow infrastructure, liability frameworks, and governance needed to act on what the tools produce.

One system's AI screening tool correctly identified atrial fibrillation risk, but the patient alert workflow was never configured — and the clinical benefit was zero.

AI tends to produce value when it targets an existing bottleneck. Without free capacity to absorb the output, AI-generated lists create frustration rather than improvement.

Systems that invested early in analytics infrastructure are compounding returns. Those that delayed are still assembling the foundation, and several leaders described the gap as widening.

One system's analytics platform, built incrementally over more than a decade, now supports a 24/7 coordination center, an ASC, an integrated practice unit, and remote monitoring — none of which would have been model-able without the underlying infrastructure.

Clinician-led analytics tend to produce faster behavior change. When the data function reports only to administrators, bedside adoption stalls.

Several leaders described absorbing growing cardiovascular demand through operational redesign rather than building additional physical capacity — though results vary by lever and maturity.

One system deploys three virtual cardiologists serving areas where in-person recruitment isn't feasible, with 80% patient satisfaction, building it into five-year plans rather than treating it as provisional.

Heart failure is the leading driver of bed pressure and ED congestion, but compensation models constrain scaling — and the CMS Ambulatory Specialist Model launching in January 2027 may force the redesign conversation regardless.

Cancer therapies are advancing faster than most systems can staff, build, or support them, leaving leaders to design buildings and care pathways with little precedent.

Roughly one-third of recent FDA approvals were in oncology, while the average community oncologist's panel is about 2% rare cancers — a structural knowledge gap that's growing.

Several leaders described building infusion centers while questioning whether home infusion, oral oncolytics, and site-neutral payment could make those investments less relevant within five years. Flexible, repurposable design is the pragmatic hedge.

Physician recruitment was the most frequently cited barrier in the opening session, but the solutions gaining traction center on rethinking how clinicians spend their time rather than simply adding headcount.

One system's APP-first and e-consult model raised the oncology share of new patient visits from 37% to 60%, creating new cancer capacity without new physician hires.

A classical hematology e-consult platform served over 1,700 patients across 30-plus locations, with roughly 88% resolved without an in-person visit — saving approximately 1.75 FTEs.

Population-scale genomic programs are expanding, but adoption depends less on the science than on reducing workflow friction and managing the trust obligations that come with custody of patients' most sensitive data.

Fewer than 10% of eligible patients are being tested. The barriers are operational — complicated ordering, hard-to-interpret results, discontinuous workflows — and defaults and automation outperform education campaigns according to one presenter.

High-profile data breaches caused measurable dips in sequencing program sign-ups, making consent and trust real constraints on enrollment.

Multiple systems are pushing expert care closer to patients through hub-and-spoke models and virtual consults, but adding locations without cultural and operational scaffolding tends to produce misalignment rather than access.

Academic partnerships tend to succeed or fail on culture, not contract structure. One system led with faculty appointments and cultural integration from the start; another described a decade-long joint venture with an unsolved pro forma.

One system's virtual consult model reached three to four times as many patients as in-person visits, with improved outcomes when subspecialty touchpoints were maintained.

Systems are making large capital commitments against uncertain timelines, and no system at the forum reported a mature, margin-positive ambulatory infusion model.

The financial case is payer-mix-dependent: moving procedures to ambulatory settings can cost traditional Medicare patients more out of pocket, while commercial populations may benefit now.

No system reported a functioning population-health model for oncology, with payer engagement described as the primary bottleneck. Several leaders recommended scenario planning over committing to a single bet.