SoundHound AI arrived at HIMSS26 with a mission that sets it apart from much of the ambient AI conversation dominating conference floors: while most healthcare AI vendors are focused on what happens inside the clinical encounter, SoundHound is targeting everything that happens around it. At the center of its case was a two-year deployment at the Medical University of South Carolina that has become one of the most detailed enterprise voice AI case studies in academic medicine.
HOTSPOT TAKE
SoundHound’s Amelia platform has moved well beyond voice transcription — MUSC Health’s 2.2-million-call deployment across patient access, revenue cycle, and pharmacy is the enterprise proof point the market has been waiting for.
SoundHound’s Enterprise Voice AI Platform
SoundHound AI has been active in voice AI for nearly 20 years, with healthcare now a primary vertical focus. Its Amelia platform combines proprietary Speech-to-Meaning architecture (which processes speech and intent simultaneously rather than sequentially) with agentic AI capabilities designed for the complex, regulated workflows of health systems. The result, according to Jeff McCool, AVP of Healthcare at SoundHound, is a platform built to handle not just simple transactions but the kind of multi-topic, conversational calls that have historically required human agents.
“So think of HR, or IT Service Desk, or even some groups that are leveraging this with their health plan that’s connected to the hospital,” McCool said, describing the enterprise-wide scope SoundHound demonstrated at HIMSS26. “We’re really trying to show the enterprise-wide capabilities of our agents.”
The platform integrates with Epic and Oracle Health (formerly Cerner), with a MEDITECH partnership in development, and is HIPAA-compliant across voice and chat channels. According to the company, its healthcare client base now spans several dozen health systems and is growing by multiple new partners each quarter. SoundHound has also expanded into adjacent use cases, including bedside meal ordering, an area it describes as generating thousands of daily calls to health system contact centers that are prime candidates for automation.
MUSC Health: From 92 Phone Lines to Statewide
The most detailed illustration of Amelia’s capabilities at HIMSS26 came from Crystal Broj, Enterprise Chief Digital Transformation Officer at MUSC Health, who joined McCool during our interview to walk through MUSC Health’s deployment of Emily — MUSC’s custom name for Amelia — across an expanding range of enterprise functions.
MUSC Health is one of the largest academic health systems in the Southeast, spanning 16 hospitals across all 46 counties in South Carolina. When the system first deployed Emily to its patient access center, the contact center was operating 30 people below capacity, wait times consistently exceeded service level agreements, and staffing shortages showed no signs of easing.
The deployment began deliberately small: appointment confirmation, cancellation, directions, and general hospital information. As the system demonstrated reliability, use cases expanded to rescheduling, Spanish-language support, and eventually full coverage across all 92 patient access phone lines. The rollout has since extended statewide.
“We started very small, with very small use cases, just to confirm, cancel an appointment,” Broj said. “And then as it became more competent, we felt more comfortable with it, and we let it reschedule appointments. Then we let her talk in Spanish, and then we extended it across all 92 phone lines.”
According to SoundHound, Emily has now handled more than 2.2 million calls at MUSC Health, with 1 in 4 fully automated without human intervention and approximately 1,000 calls handled overnight without staff involvement. Patient satisfaction with the agent runs 4.4 out of 5.
Results Across Revenue Cycle and Pharmacy
The patient access success led MUSC to extend Emily into revenue cycle management, where agent burnout is a persistent operational challenge. Billing inquiries, balance checks, and payment method guidance now achieve a 25-to-30 percent call deflection rate. Pharmacy followed, handling prescription refill questions and medication inquiries that previously required pharmacist time. All deployments integrate with MUSC’s Epic EHR system.
The workforce impact has been notable. Rather than eliminating positions, the platform has allowed MUSC to absorb growth and close its staffing gap without proportional headcount increases.
“We were down 30 people in the contact center just because we could not hire. And so our staff was overburdened, and we had long wait times, and we weren’t meeting our service level agreements, and now we are.” — Crystal Broj, Enterprise Chief Digital Transformation Officer, MUSC Health
“We were down 30 people in the contact center just because we could not hire,” Broj said. “And so our staff was overburdened, and we had long wait times, and we weren’t meeting our service level agreements, and now we are.”
McCool framed the growth dimension from SoundHound’s perspective: “They have been able to acquire and grow, and not necessarily need to hire more and more agents to achieve that growth.”
Employee satisfaction has followed. Pharmacy staff, Broj noted, are “over the moon” with the platform’s efficiency. Press Ganey scores in the “ability to contact my provider” category have improved as hold times have dropped for patients reaching human agents. The agents themselves, once fielding an unrelenting volume of routine calls, now have the capacity for more complex patient coordination that genuinely requires human judgment.
“You want the agent to work on ‘is my appointment at one or at two [o’clock],'” Broj said. “Emily handles those calls now.”
Agentic AI: SoundHound’s Next Technology Layer
HIMSS26 also allowed SoundHound to articulate its transition from deterministic conversational AI to a fully agentic framework, a shift McCool described as a natural evolution for health systems that have already proven the value of the underlying platform.
Pharmacy at MUSC is already running on an agentic framework, according to McCool, providing early production evidence for the approach. Facilities management and HR are next: Emily will generate maintenance tickets from staff calls and, through a Workday integration, enable employees to check benefits, schedule HR consultations, or explore job openings via voice or chat from a mobile device.
On the technology roadmap, SoundHound is developing a streaming voice engine for MUSC that enables real-time, interruptible conversation, eliminating the turn-based interaction model of earlier systems. Existing deterministic workflows for functions like authentication will remain in place while the agentic layer opens up for post-authentication interactions.
“Our goal is to transition them to our streaming voice engine. So right now, just like you and I are talking, you don’t need to wait till I finish this sentence before you understand what I’m saying. That’s where we want to take Amelia next.” — Jeff McCool, AVP of Healthcare, SoundHound
“Our goal is to transition them to our streaming voice engine,” McCool said. “So right now, just like you and I are talking, you don’t need to wait till I finish this sentence before you understand what I’m saying. That’s where we want to take Amelia next.”
Competitive Position in a Crowded Market
SoundHound’s healthcare voice AI strategy occupies a distinct layer of the clinical enterprise: the patient-facing and internal operational call volume that falls entirely outside the physician encounter. That positions Amelia alongside, rather than against, ambient documentation tools like Nuance’s Dragon Copilot or Heidi, which focus on the clinical note.
Within its own competitive space, SoundHound faces challengers including Hyro and Syllable, which target patient-facing call automation in ambulatory settings. SoundHound’s differentiation rests on the Speech-to-Meaning architecture’s ability to handle multi-intent conversations without losing context, and on the depth and breadth of its enterprise deployments. MUSC Health’s multi-department rollout, alongside Allina Health’s “Alli” deployment in 2025, provides the kind of named, measurable reference cases that health system procurement committees increasingly require.
“We looked at three or four different vendors, but we chose SoundHound for the quality of the types of calls that it could handle,” Broj said. “I could say, ‘I want to schedule an appointment, and I need to pay my bill.’ And Emily will come back, ‘Okay, let’s talk about rescheduling your appointment,’ and then, ‘Oh, and did you want to pay your bill now?’ And she can switch that mode and remember what you wanted to talk about, and then go and do that action for you.”
For patients navigating a health system that spans the entirety of South Carolina, that capability is not a feature demonstration. It is the access point.
— This original article was created with AI support.
