Voice Agents Setup and LLM Configuration Improvements
Release date: May 7–8, 2026
Release type: Bundled enhancement
Product area: Voice Agents
Voice Agents setup and configuration improvements
We released a set of improvements that make Voice Agents easier to find, duplicate, configure, and test. The update includes clearer Voice Agents naming, a new clone workflow, voice previews, improved duration editing, refreshed advanced configuration controls, and a cleaner LLM model picker/default model flow.
What’s new
Clearer Voice Agents navigation and labels — The product now uses more specific Voice Agents naming across the relevant navigation and setup areas, making it easier for users to identify the right workspace.
Clone Agent workflow — Users can duplicate an existing Voice Agent’s core setup instead of rebuilding the same configuration from scratch.
Voice previews — Users can preview available voices from the voice picker before saving or testing an agent.
Improved Max Duration editing — The Max Duration field now behaves more naturally when users clear and retype values.
Advanced Settings improvements — Advanced controls are easier to access, while LLM-specific configuration can be shown only where enabled.
Refreshed LLM picker — The model picker is cleaner and easier to scan, with updated model metadata, grouping, labels, and default model behavior.
Customer impact
Faster setup for customers and Ops teams creating similar Voice Agents.
Less trial-and-error when choosing voices.
Clearer navigation and terminology for Voice Agent workflows.
Easier editing when adjusting Voice Agent duration settings.
Better control over which organizations can access LLM-specific configuration.
More consistent default behavior when creating or updating Voice Agents.
Where to find it
Users can find these improvements in the Voice Agents setup and edit experience, including the Voice Agent list, create/edit forms, voice picker, advanced settings, and LLM configuration controls.
Notes
No customer action is required for the naming, picker, and form behavior improvements.
Access to LLM-specific configuration may depend on organization-level enablement.
Cloning copies the core Voice Agent setup, but some connected resources may still need to be configured separately after cloning.