Southwest Medical Imaging (SMIL) provides state-of-the-art diagnostic medical imaging and interventional radiology through 17 outpatient centers. More than 50 radiology physicians, each fellowship-trained in one of 10 imaging specialties, serve thousands of patients who count on SMIL for reliable, easy-to-manage imaging care. The organization employs around 670 people, and its contact center handles roughly 6,500 calls a month, with appointment management among the most common request types.

Before working with Replicant, SMIL's contact center faced the pressures common to the industry. Call volumes were high, hiring and retaining agents was difficult, and the agents it did have lacked time for the requests that most needed a person. The center's answer rate stood at 73%, and its abandonment rate hovered between 19% and 30%, which meant long hold times for patients.

Appointment management was one of the most frequent reasons patients called, and much of it was routine: cancelling an exam, confirming a time or asking a scheduling question. Cancellations were a manual process, since an agent had to go into the scheduling system and remove the appointment time by hand. Until that happened the slot did not show as available, so it could not be offered to another patient and the revenue went unrealized.

When SMIL's leadership discovered customer service AI, they saw its potential to resolve several appointment-related use cases at once. Having already spent time with another vendor, the team began exploring how Replicant's Thinking Machine could raise the answer rate, reduce abandonment and give live agents room to focus on patients who needed help with harder issues.

Replicant designed a Thinking Machine, its conversational AI agent, to fully automate appointment cancellations and confirmations over the phone. Replicant integrated the system with SMIL's existing technology stack, so when a patient cancels, the appointment time opens up immediately, whatever the day or hour of the call. Confirmations are handled the same way, without an agent touching the scheduling system. The initial use cases covered exam cancellations, appointment confirmations and appointment FAQs.

SMIL launched Replicant early and strategically, giving agents clarity on how automation would improve their experience. According to Kim Wyszynski, SMIL's Director of Contact Center, Replicant was extremely responsive from the start and able to move quickly: meetings happened as scheduled and were followed up immediately.

Once live, patients became accustomed to managing appointments over the phone with the Thinking Machine, and agent workloads decreased. Replicant's Customer Success Managers and SMIL's care team continue to work in lockstep to design conversation flows and build out new use cases, with the goal of making every patient experience more patient-centric.

• Average answer rate for agents climbed from about 73% to over 90%
• Call abandonment fell from around 25% to as low as 5% on average
• Replicant takes up to 700 calls a day, removing repetitive requests from agents' queues
• Cancelled slots open in real time, so more appointments are filled and less revenue is lost
• Agents often have one day a week off the phones to learn new skills and focus on strategic work
• Patients manage appointments at any hour without holding, and urgent callers reach an agent faster