A Medical Imaging Leader's Contact Center Overhaul on AWS

It's 8am on a Monday, and a patient is calling to reschedule an MRI. She punches through a maze of prompts, gets transferred twice because the system can't tell who she is or why she's calling, and repeats her date of birth to a third agent before anyone can actually help her. By the time she hangs up, she's spent longer navigating the phone tree than she will in the MRI.

Now multiply that by a network of hundreds of imaging centers nationwide, and the pattern becomes impossible to ignore: 

• A legacy, keypad-only IVR driving long, frustrating calls
• No reliable way to identify or route callers, so transfers pile up
• Agents working without a unified view of the patient or the case, that causes longer training and inconsistent service

What changed

Rather than a simple upgrade, this healthcare provider partnered with NeuraFlash and AWS to modernize its contact center from the ground up. NeuraFlash implemented Amazon Connect with Amazon Lex for speech-enabled IVR, real-time intent recognition, and intelligent routing, paired with NeuraFlash Desktop to unify the agent workspace with guided flows, screen pops, and real-time analytics. Contact Center Conversational Analytics (formerly Contact Lens), Amazon Bedrock, and Amazon Q in Connect were layered on top to bring agentic service to both patients and agents.

Here's what their modernization unlocked:

• Authenticated calls 1.5 minutes faster, through automated identity verification and upfront data capture 
• A full week shaved off agent training, thanks to guided workflows and a unified desktop that gets new hires up to speed quicker
• 23 seconds saved per call, from automated call summaries and streamlined wrap-up
• A ready foundation for agentic AI, with Bedrock and Q in Connect layered onto Amazon Connect

It doesn't stop there

The modernization gave this provider a stable, AWS-native foundation and NeuraFlash is now building on it toward something more ambitious: a fully agentic scheduling experience. Instead of just capturing a caller's request for a human to act on later, the next phase puts a conversational agent, built on Amazon Lex and Amazon Nova Sonic for natural, speech-to-speech interaction, directly in the loop. Patients will be able to create, reschedule, or cancel an appointment in English or Spanish, in their own words, no keypad required. Behind the scenes, Amazon AgentCore gives that agent secure, governed access to the provider's own scheduling systems, so it isn't just capturing a request, it's resolving it.

It's a playbook NeuraFlash has now repeated across 300+ Amazon Connect implementations: bring the stack together, take the manual work off agents' and patients' plates, and leave both care teams and compliance with something they can depend on. We don't build it and walk away. We stay in it with you, continuously optimizing as your needs evolve and the technology moves forward.

Ready to see what this could look like for your organization? Book a discovery call with us today.

‍