I recently had the privilege of joining senior leaders from Johns Hopkins, Mayo Clinic, and USF Tampa General on a Becker’s Healthcare panel focused on unlocking quick AI wins. The conversation was candid and the ambitions for AI were high. We briefly touched on clinical AI and ambient listening. In terms of business operations, one theme cut through everything else: the health systems seeing results right now are the ones that started where the patient actually is…on the phone. That is where advanced AI delivers its fastest, most measurable win.
Specifically, it was the moment Matt Cantonis, COO of USF Tampa General Physicians, described what happened five weeks after deploying an AI-powered front end to their phone system. With the same scheduling staff and the same number of agents, the department made 20% more appointments in the very first week. It wasn’t because they hired more people. It wasn’t because they redesigned the patient portal to be more user friendly. It’s because they stopped wasting human capacity navigating healthcare callers through a system that a voice AI agent could handle in seconds.
That is the quick win most health systems are not taking seriously enough.
The AI Conversation Has a Blind Spot
The majority of healthcare executives say they are actively investing in AI to streamline administrative operations. The majority of that conversation, understandably, gravitates toward ambitious long-term initiatives. But the Becker’s panel illustrated a pattern that I encounter constantly: organizations are so focused on the transformative horizon that they are leaving substantial, immediate operational value on the table.
Voice channel automation (AI-driven patient care navigation and call routing) is consistently the fastest, lowest-risk path to measurable ROI in healthcare operations. And yet, as I noted during the panel, many health systems are still running 50-year-old push-button phone trees. Patients wade through queues to reach agents who then transfer them elsewhere, where the process starts over. One contact center I spoke with recently handles only 30% of inbound call traffic. The other 70% navigates a system that was never designed for them.
This is not a technology problem. It is a prioritization problem.
What “Quick Win” Actually Means
Matt’s results were not an anomaly. They were a confirmation of what Parlance has seen in thousands of hospital and clinic deployments, demonstrated over the past three decades. Parlance client, Alina Health, a mid-sized health system with 11 hospitals and 68 clinics, after deploying a conversational AI platform, documented 6,070 agent hours saved per month and $1.45 million in annualized savings. Another Parlance client, UC San Diego Health, who was managing 1.5 million calls annually eliminated the equivalent of nearly 18 weeks of human labor within the first month — before a single new hire was made or a single workflow was redesigned. Over and over, we see concrete evidence that simplifying caller interactions delivers clear operational and financial returns.
The mechanism is straightforward. 84% of healthcare practices report that patients use the phone to schedule care. While digital adoption is growing, phone calls remain the “front door” for the majority of healthcare interactions, particularly for complex tasks like rescheduling or clarifying medical questions. Despite billions invested in digital portals, only approximately 10% of appointments are booked online nationally. Leading health systems report that the average patient makes 3.5 calls for each single scheduling need they have, regardless of the booking channel. The volume is not declining. The staff capacity to absorb it is.
When AI handles call navigation, routing, patient verification, and FAQ response, agents are freed to operate at the top of their capabilities — managing complex access coordination, scheduling nuanced cases, and delivering the empathetic interactions that actually build patient loyalty. The math is direct: reduce average handle time on routine calls, and the same team delivers materially more productive output.
One Parlance client, Virtua Health, with $3.2 billion in net revenue and 8,500 employees reported a 13% increase in patient-rated ease of contact and a 29% improvement in ease of scheduling over nine months. Measured by Press Ganey, these aren’t just “feel-good” numbers, they are the critical satisfaction signals that secure patient retention in an increasingly competitive healthcare market.
Starting Small Is Not a Weakness. It Is the Strategy.
Maneesh Goyal of Mayo Clinic made a point during the panel that deserves emphasis: full agentic scheduling automation, where AI dynamically books appointments across complex, multi-specialty workflows without human intervention, may be years away for most organizations. The data standardization, EHR readiness, and provider-level scheduling logic required to get there are not trivial problems. These are complicated logistics… Even if it is technically possible sooner than we can imagine, there is still the constraint of doctors’ having the appointment availability. Will there be a day when a patient with cancer can use AI to help coordinate appointments for bloodwork, a cancer doctor, maybe another specialist, etc., and schedule all the appointments on a single day, and not have that day be 2 years away? This is an interesting problem to consider… But that future state is not the prerequisite for acting now.
The entry point is not finding appointment slots or navigating the full complexity of orchestrating multi-specialty provider scheduling at the edge of what is technically possible. The entry point is navigating healthcare callers efficiently, ensuring that every patient who calls reaches the right destination the first time, without unnecessary transfers, without hold time, without misdirection. When the patient has a frictionless experience, they feel in control of their care journey, and less overwhelmed by the complexities of the healthcare system.
Modernizing call navigation is the sensible starting place. From that foundation, health systems layer in self-service capabilities incrementally: appointment rescheduling for established patients with existing providers, patient verification, FAQ automation, and eventually more complex scheduling workflows as EHR standardization matures. This is not a compromise. It is the architecture of sustainable, compounding ROI.
The Governance Argument for Starting Here
Eighty-five percent of health systems are using AI internally. Only 17% have mature governance structures.* The panel’s consensus on governance was unambiguous: the organizations moving fastest and most safely are those that have defined the rules of engagement before scaling.
Voice channel automation presents an ideal governance on-ramp for exactly this reason. The risk profile is narrow, the use case is well-understood, the ROI is measurable from day one, and HIPAA compliance requirements are non-negotiable and manageable. A health system that deploys conversational AI in the contact center is not taking a speculative bet on probabilistic technology. It is operationalizing a proven solution with three decades of implementation evidence behind it.
The financial return is real. The governance muscle developed through the implementation applies directly to more complex AI initiatives that follow.
Elevating People, Not Replacing Them
One of the most resonant segments of the panel discussion centered around contact center staff member reactions to voice AI agent deployment. Rather than experiencing the intervention as a threat, teams found greater purpose and their patient interactions became more meaningful because the low-value volume had been absorbed by the system.
This is consistently what the data shows. Virtua Health, a $2.8 billion New Jersey system, achieved a 92% employee retention rate in their contact center following voice AI implementation, with a median tenure of 3.9 years, well above industry averages. When agents are no longer fielding parking directions and repeat transfer requests, they perform better and stay longer.
The labor math works in both directions. Voice AI does not simply reduce headcount. In most deployments, it eliminates the need to grow headcount as call volume increases, which, given persistent staffing shortages across non-clinical healthcare roles, is often the more valuable outcome. US News and World Report found that 44% of non-clinical healthcare staff report being overloaded, and approximately a third intend to leave their positions within two years. Automation that reduces that pressure is not a headcount elimination strategy. It is a workforce sustainability strategy.
The Right Framework for Measuring ROI
The panel surfaced an important tension in how health systems evaluate AI investments: it is genuinely difficult to isolate the contribution of any single technology when multiple operational initiatives are running in parallel. Matt acknowledged this directly, noting the concurrent work on scheduling templates and call routing decisions.
For voice channel automation, the isolation problem is more tractable than most AI use cases. The metrics are discrete: agent hours saved, call abandonment rate, first-call resolution rate, appointment conversion rate, patient satisfaction scores on accessibility and ease of scheduling. These are established KPIs with clear baselines. The before-and-after delta is attributable.
The deeper ROI framework, as Maneesh articulated, ultimately traces to patient satisfaction and bottom-line performance. Voice automation, and specifically the quality of patient care navigation at the point of first contact, earns its place on both dimensions: directly through labor savings and appointment volume, and indirectly through the patient experience improvements that drive retention in an environment where the cost of losing a patient far exceeds the cost of serving them well.
The Imperative to Move
My closing message to the Becker’s audience was the same as it is here: keep moving forward. The health systems producing results are the ones deploying, measuring, learning, and iterating — not the ones waiting for perfect conditions.
Almost any health system is able to go from first vendor conversation to live deployment in a few months. The results will be visible within the first week. That is the velocity available to health systems that treat voice channel automation as the operational priority it is — not a future-state aspiration, but a present-state financial lever available today.
The phone remains the highest-volume patient interaction channel in every health system in the country. The technology to modernize it is proven, the ROI is guaranteed, and the governance risk is manageable. The only question is how long operational leaders are willing to let a solved problem continue to cost them.
*As of August 2025
Scott D’Entremont is CEO of Parlance, a HIPAA-compliant omni-channel AI platform built specifically for healthcare and backed by Constellation Software. Parlance serves hundreds of health systems across voice, SMS, and web chat, delivering measurable ROI within 30 days. Scott participated as a panelist in the Becker’s Healthcare webinar “Unlock the Power of Quick AI Wins” alongside senior leaders from Johns Hopkins, Mayo Clinic, and USF Tampa General.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.