CTG’s Esdale and Kochan say Patient Contact Centers are the Overlooked Frontier for AI Optimization

Podcast: Play in new window | Download (Duration: 24:54 — 28.5MB)

Subscribe: Apple Podcasts | Spotify

Patient contact centers at health systems handle 40 to 50 times more calls per day than IT service desks/contact centers, yet many organizations still treat them as fixed operational costs with little room for strategic improvement. In this interview from the ViVE conference, part of our Partner Perspective series, CTG’s Joseph Esdale, Strategic Solutions Partner, and Christina Kochan, Healthcare Solution Architect, discuss how agentic AI and modern contact center platforms can transform the patient calling experience, reduce workflow friction, free clinical staff from documentation burdens, and scale across entire organizations.

Health systems that view their patient contact centers as little more than a cost of doing business are sitting on one of the largest untapped opportunities for AI-driven improvement in their organizations.

That’s the assessment from Joseph Esdale, Strategic Solutions Partner, and Christina Kochan, Healthcare Solution Architect, at CTG, a company that has supported more than 1,000 healthcare organizations over 35 years with EHR implementations, managed services, cybersecurity, and clinical optimization. In a recent interview, the pair described how agentic AI is reshaping the service desk landscape for health systems, starting with a function many CIOs have historically deprioritized.

Podcast: Play in new window | Download (Duration: 24:54 — 28.5MB)

Subscribe: Apple Podcasts | Spotify

The Overlooked Giant

The math tells the story. Patient contact centers at large health systems typically handle 40 to 50 times more calls per day than the IT service desk. Yet the IT service desk has traditionally received more strategic attention, while patient-facing call centers have been managed as an operational necessity. “A good CIO friend of mine calls anything around the service desk, “sunk costs that you just have to pay,” Kochan  said. “I think it’s been largely left unrecognized as a place that you can optimize.”

That calculus is shifting. Agentic AI platforms can now handle complex patient interactions: scheduling appointments, answering questions about test results, processing bill payments, and routing callers to the right department with full context already in hand. The technology eliminates one of the most persistent patient pain points, the need to repeat personal information and restart the conversation with every transfer. Modern systems validate a patient’s identity up front and pass that context forward, so each subsequent agent already knows who is calling and why.

On the clinical side, AI-powered transcription and call summarization are addressing a significant documentation burden. Clinicians working triage lines often spend up to 15 minutes on a call with a patient and another 8 to 10 minutes documenting the conversation in the EHR. Modern contact center tools generate a transcript and call summary within 30 seconds of the call ending, ready to post into the patient’s record. CIOs are well positioned to identify these systems and bring them to the attention of clinical leaders who may not have the bandwidth for a thorough product evaluation.

Start With the Patient’s Experience

The most revealing diagnostic for any service desk operation is deceptively simple: follow the patient through the entire process. A personal example from Kochan illustrated the point. After developing back issues, she received a digital scheduling ticket from her healthcare provider to book an MRI. The ticket’s instructions: call the office to make the appointment. She called, only to learn the prior authorization had not been completed. The appointment could not be scheduled, so she deferred. “I know the system. I know how it works, and I still can’t get my MRI scheduled,” Kochan said.

That kind of friction exists across healthcare organizations, often because patient-facing workflows span multiple departments, systems, and external entities like payers. A provider-centric workflow analysis captures the operational view but frequently misses the patient’s actual experience. Mapping each step from the patient’s perspective exposes breakdowns that are invisible from the inside. Generative AI tools embedded in modern contact center platforms can analyze call data to reveal exactly what patients are calling about and in what volume, replacing the guesswork many organizations still rely on. “With modern tools, you can get generative AI that tells you exactly what patients are calling about,” Esdale said. “As you start building a roadmap of how you’re going to modernize, you can understand the low-hanging fruit.”

The goal is to address friction points iteratively, targeting the highest-volume issues first. One pass will not fix everything; the data from generative AI call analysis allows teams to plan successive rounds of improvement with clear priorities.

Choosing Platforms Over Point Solutions

For CIOs evaluating AI voice agents, the future proofing question is critical. A standalone AI agent that handles one function for one department creates the same fragmentation that plagues legacy contact center operations. A platform approach allows health systems to build and deploy new agents, integrate across clinical and administrative workflows, and handle surges in demand during events like natural disasters. Kochan emphasized the importance of selecting vendors whose platforms can shift and flex as organizational needs evolve, supporting new agent configurations at scale.

CTG recently demonstrated this model in a proposal for a poison control center. The end-to-end solution paired Amazon Connect as the phone system with Hippocratic AI as the agentic AI layer. Amazon Connect’s AI tools surface relevant reference articles for human agents in real time, while Hippocratic AI’s agents conduct clinical conversations, gathering patient information and triaging appropriately. The human-in-the-loop element remains essential for clinical decision-making; AI agents can identify unsafe situations and escalate to clinicians, but treatment decisions stay with licensed practitioners.

The scaling question extends beyond call volume. CIOs frequently contend with departments that adopt niche vendor solutions for radiology, oncology, or other specialties, each requiring separate EHR interfaces and consuming IT resources. Selecting a platform vendor capable of serving all departments through a single integration model reduces that complexity. Governance is equally important; a board overseeing the entire patient contact experience prevents the fragmented approach where scheduling, billing, and medical records each operate their own call center with no shared context between them.

Take it Away

• Patient contact centers handle 40 to 50 times more calls than IT service desks and represent the bigger optimization opportunity for AI deployment
• Map the patient’s calling experience end to end before selecting or evaluating AI tools; provider-centric workflow analysis misses critical friction points
• Use generative AI to categorize call types by volume and prioritize improvements iteratively
• Select agentic AI platforms that can scale across departments and adapt to new use cases
• Establish a governance board over the entire patient contact experience to prevent fragmented call center operations
• AI-powered transcription and call summarization can cut clinical documentation time nearly in half after triage calls
• Human-in-the-loop remains non-negotiable for clinical decisions; build escalation pathways into every AI deployment

For CIOs weighing where to direct their focus, Kochan offered a simple guide. “Getting in there and following the patient’s experience is going to take you pretty far in optimization,” he said. “I think CIOs need to know that.”