How does MUSC use AI with patients?

An AI agent called Emily, built on SoundHound and integrated with Epic, handles routine calls and chats such as appointment changes, billing and refills.

Can a small practice do this?

Start smaller. List your front desk’s most common non-clinical questions and write clear answers. That becomes the knowledge base for any phone or chat tool.

Keep patient data out of general-purpose AI tools unless your vendor signs the right agreements. Use AI for non-clinical, non-identifying work first.

Where do the numbers come from?

From SoundHound’s published case study. They are vendor-reported.

Directory / AI for Healthcare

AI for healthcare in Charleston

The biggest healthcare AI example in Charleston is MUSC Health. Its AI agent, Emily, handles appointments, billing questions and pharmacy refills by voice and chat, and SoundHound reports 2.2 million interactions a year. Smaller practices can start with the same idea: answer the routine questions automatically and send the rest to staff.

Updated October 16, 2026

Local examples

01MUSC HealthCharlestonSoundHound AI agent "Emily", integrated with Epic2.2M interactions/yr; 4.3/5 NPS; billing abandonment down up to 40% (vendor story)
02Roper St. Francis Healthcare (Palmetto Market at Bon Secours St. Francis Hospital)CharlestonAmazon Just Walk OutPartner announcement: called the first full-service coffee shop with barista service in a U.S. healthcare facility to use this technology, with no checkout line

The AI news that matters for your small business, explained simply, with how to use it. Free weekly, from Charleston.

Common questions

How does MUSC use AI with patients?: An AI agent called Emily, built on SoundHound and integrated with Epic, handles routine calls and chats such as appointment changes, billing and refills.
Can a small practice do this?: Start smaller. List your front desk’s most common non-clinical questions and write clear answers. That becomes the knowledge base for any phone or chat tool.
What about privacy?: Keep patient data out of general-purpose AI tools unless your vendor signs the right agreements. Use AI for non-clinical, non-identifying work first.
Where do the numbers come from?: From SoundHound’s published case study. They are vendor-reported.