ElevenLabs Adds Conversation Triage Tickets and Twilio Answering Machine Detection to Agents Platform

ElevenLabs shipped a batch of ElevenAgents platform updates on September 7, 2026, adding conversation triage tooling, dynamic filtering, and answering-machine detection for outbound calls — features aimed squarely at teams running AI voice agents for customer support and outbound calling at scale.

Details

• Conversation triage tickets: a new endpoint retrieves conversation triage tickets across all agents in an account, with filtering by status or assignee and cursor-based pagination for up to 100 tickets per page
• Dynamic variable filters: conversation listing and message search now support repeatable name:op:value filters, including equality and comparison operators (eq, gt, gte, lt, lte) on numeric values
• Twilio answering machine detection: outbound calls can now enable machine detection in two modes — an immediate-verdict enable mode or a detect_message_end mode that waits for voicemail greetings to finish — with results delivered through new webhook events
• Agent metadata and knowledge base updates: branch responses now include draft creation timestamps and version tracking; RAG query chunks now require source_url fields; WebSocket agent_response events support optional message attachments with URL, name, and MIME type
• Data collection changes: properties now use an allowed_values field referencing dynamic variables, deprecating the previous field name
• SDK and CLI updates: JavaScript, Python, and CLI SDKs received version bumps for dynamic variable support and regenerated types; the CLI adds an elevenlabs say command and now defaults to the eleven_v3 TTS model

What happened next

The updates are live now across ElevenLabs’ API, SDKs, and CLI as documented in its changelog. Answering machine detection in particular addresses a long-standing pain point for outbound voice-agent deployments — distinguishing a live person from voicemail — while the new ticketing and filtering tools point to ElevenLabs pushing ElevenAgents further into enterprise contact-center territory rather than treating it purely as a conversational-AI demo layer.