Home
Contact usLog in
Release notes

Webhook Tool Filler Messages Now Available for Telnyx AI Assistants

6, Jul 2026

AI Assistants can now speak scripted utterances at the start of a webhook tool call and during long-running requests. Callers stay engaged while the assistant waits on an external API, database lookup, or third-party model, instead of sitting in silence.

What's new

  • Filter messages on webhook tools: Set the exact phrase the assistant speaks when a synchronous webhook starts, and tier additional phrases at delay thresholds (5s, 10s, 15s) while the call is still executing.
  • Deterministic utterances: Messages are scripted per tool, not LLM-improvised on the fly. You control wording, timing, and brand voice for every wait state.
  • Sync webhook and MCP tool support: Filler messages apply to synchronous webhook tool calls and MCP tool calls where the assistant holds the line until the response returns. Async webhooks are not supported.

Why it matters

Silence is where voice AI calls die. A webhook tool call that takes 8 to 12 seconds for a database lookup or third-party API leaves the caller staring at dead air, and on a phone call that reads as broken, not loading. Most voice platforms either go silent and lose the caller, or ask the LLM to improvise filler, which rambles, drifts off-brand, and can't be tuned.

This lets you script the wait-state UX the same way you script the conversation flow. "Let me look that up for you" at the start, "still working on this" at 5 seconds, "almost there" at 15 seconds. Deterministic, tiered, and per-tool. The same control you have over the conversation now extends to the gaps the conversation depends on.

Example use cases

  • Customer support agents that query a billing or CRM system mid-call can acknowledge the lookup and reassure the caller instead of going mute.
  • AI receptionists booking appointments through a scheduling webhook keep the caller on the line with timed acknowledgments while the calendar API responds.
  • Voice agents calling slow third-party LLMs or RAG pipelines can set longer delay tiers to match realistic execution times without callers assuming the call dropped.

Getting started

  1. In Mission Control, open AI, then Assistants, and select your assistant.
  2. Go to the Webhook Tools section and edit or create a webhook tool.
  3. Select Sync as the Request Mode.
  4. Open the Filter Messages tab.
  5. Add a message for "When the request starts" to set the opening utterance.
  6. Add delay-tiered messages with a delay in milliseconds for each mid-call phrase.
  7. Save the tool.

Each filler message has a type of request_start (spoken immediately when the tool call begins) or request_response_delayed (spoken after timing_ms milliseconds if the response has not yet arrived). The timing_ms field accepts 100 to 120000 ms and is only required for delayed messages. Once the response arrives, any remaining delayed messages are skipped.

{
  "filler_messages": [
    { "type": "request_start", "content": "Let me look that up for you." },
    { "type": "request_response_delayed", "content": "Still working on this.", "timing_ms": 5000 },
    { "type": "request_response_delayed", "content": "Almost there.", "timing_ms": 15000 }
  ]
}

Learn more in the AI Assistants docs or the Telnyx docs.