iCreatAI News
Skip to content
Trending storyUpdating

ElevenLabs releases tutorial on building a healthcare outbound reminder Agent

1 reports1 sources8 hr agoUpdated

Understand this story

AI overview

ElevenLabs has released a tutorial showing how to build an outbound voice agent on ElevenAgents using Architect: calling patients before appointments, verifying dates of birth, and during the call confirming, rescheduling, or canceling appointments within the electronic health record (EHR). The tutorial explains several usage conditions: batch outbound calls only use a portion of the concurrency quota (whichever is lower, 50% of the workspace limit or 70% of the agent limit), to avoid crowding out inbound lines; zero-retention mode can be enabled for individually initiated calls but not for batch outbound calls; however, Architect conversations are not covered by zero-retention mode. ElevenLabs states that to qualify for HIPAA, a BAA must first be signed in an enterprise agreement before real patient data can come into contact with the agent.

AI overview based on source reports · Updated 54 min ago

Coverage timeline

Follow the coverage to explore different perspectives.

Nov 20
  1. ElevenLabs Blog
    ElevenAgents for Healthcare: Build an outbound appointment reminder agent with Architect

    ElevenLabs describes building an outbound voice agent on ElevenAgents, configured via its Architect assistant, that calls patients before a visit, verifies date of birth, and confirms, reschedules, or cancels the appointment in the EHR during the call. The agent can run as a nightly batch or in real time from EHR events, and the post reports that simulation tests initially exposed failures where appointment details leaked before verification and a reschedule time was invented, which were fixed by moving details into the confirm node, rewriting node prompts, and switching models. It also notes live calls revealed repeated acknowledgements and a reschedule write that silently dropped the participant and slot arrays, caught only by reading the resource back.

Activity for this story

Not enough continuous observations to show a trend yet.