AI Call Assistant

Overview

The AI Call Assistant is a voice agent that answers WhatsApp business phone calls, holds a real spoken conversation with the caller, can look things up or take actions mid-call (its skills), and records every call in Call Logs. You build and manage assistants under More → AI Call Assistant.

You use this page to set up the assistant and review its calls. The live conversation itself — turning the caller's voice into text, sending it to the AI, turning the AI's reply into speech, and playing it back — runs automatically in the background once a call comes in. The assistant simply uses the settings you save here.

Each assistant belongs to one workspace and is built with a five-step wizard: Identity, AI Brain, Skills, Voice, and Routing. Every assistant has a status (Live, Draft, or Paused) and an on/off switch; it only answers calls when it is switched on and set to Live.

Prerequisite — the admin sets the provider keys. The AI provider keys (OpenAI, Anthropic Claude, Google Gemini) and voice keys (ElevenLabs for the spoken voice, Deepgram for transcription) are set up once by the platform admin under Admin → AI & API Keys. You do not need to paste a key — every assistant uses the admin's keys automatically (see Your own keys vs admin keys). Add a per-assistant key only if you want this one assistant billed to your own account.

Plan-gated — two features. Building assistants needs the AI Voice Agent plan feature. The calls those assistants answer, plus the Call Logs viewer, need WhatsApp Calling — the same feature behind WhatsApp Calling. Playing back call recordings needs a third feature, Call Recording. If your plan doesn't include these, the menu items are hidden.

How a live call works

Knowing what happens during a call makes every wizard field easier to understand. Each call runs through these steps:

1. WhatsApp delivers the call. An incoming call arrives for your number, tagged with the caller's details and which assistant should answer.
2. The assistant loads your settings. It pulls in the assistant's configuration and the voice keys for your workspace. If either the transcription key (Deepgram) or the speech key (ElevenLabs) is missing, the call is declined up front so the voicemail fallback handles it instead of leaving the caller in silence.
3. The call connects. The audio link to the caller is set up and answered.
4. Greeting. As soon as the line is live, the agent speaks a greeting — picked at random from your greeting variations — so the caller never hears dead air.
5. Listen. The caller's speech is transcribed to text in real time (using Deepgram). Each finished sentence becomes one transcript turn.
6. Think. The agent takes the recent conversation (roughly the last 12 turns) plus your behaviour brief and asks the AI model you chose for a reply.
7. Speak. The reply text is turned into a natural voice (using ElevenLabs) and played back to the caller.
8. Loop / exit. Listen, Think, and Speak repeat. If the caller says a hangup keyword, the agent speaks the goodbye line and ends the call. When the call ends, recordings are saved and the final transcript is stored.

Transcript turns are saved as the call happens, which is why a call in progress already shows a growing transcript in Call Logs.

Creating an assistant

From the AI Call Assistant list, click New assistant. The wizard always saves the full assistant state at once, so you can jump between steps freely and click Save at any time. Saving a brand-new assistant returns you to its edit URL so you can keep refining it.

1. Open the wizard (or open an existing assistant to edit it).
2. Work through the five steps below. The side panel shows a live, text-only preview of how a call would read, plus the chosen AI provider and voice.
3. Set the status. Leave it as Draft while building; flip the Live switch in Step 1 when you are ready for it to answer real calls.
4. Click Save. The assistant is saved to your workspace, and all skills are saved exactly as shown in the form.
5. Test from a real phone before relying on it — the in-app preview is text-only and doesn't test the voice or the actual phone call.

A brand-new assistant starts with the defaults shown in the reference tables below; you can change any of them.

Step 1 · Identity & persona

Defines who the agent is and how it opens a call.

Persona presets

Step 2 · AI Brain

Chooses which AI model powers the agent and how it behaves. You can use any of three providers — pick one, then set the exact model name, an optional key of your own, the behaviour brief, an optional knowledge-page link, and three personality sliders.

Model reference

These are the model names the wizard suggests for each provider. The default for a new voice agent is gemini-2.5-flash-lite — chosen because Gemini responds fastest on voice calls.

Behaviour brief & personality

• Behaviour brief. State who the agent is, what to ask for, which skills to use, and when to hand off to a human. If you leave it blank, the agent uses a generic instruction: "you are a helpful voice assistant on a phone call — reply with one short sentence at a time so the caller can interject."
• Knowledge source (web link). An optional page the agent can reference so it answers from your own content.
• Warmth / Formality / Pace. Fine-tune the delivery beyond a single tone setting. Pace also keeps the agent concise so it doesn't over-talk.
• Your own key. Leave blank to use the admin's key. If a key is already saved, the field shows "saved · leave blank to keep" and an empty box never erases it.

Step 3 · Skills (tools)

A skill lets the agent call out to one of your web services in the middle of a conversation — to look up an order, book a slot, or create a ticket. The agent is told "call this if the caller asks about X"; when it decides to, it sends a request to your web address with the details it picked up from the conversation, then uses the response in its reply. Each assistant can have up to 25 skills.

Skills can chain: the agent calls one, reads the response, then calls another in the same turn. You can use @placeholder tokens in skill values that are auto-filled from the caller's history. Every skill the agent uses is recorded in the tool-call timeline with its details and the time into the call.

Save behaviour. Saving stores the skills exactly as shown in the form — so removing a skill in the form removes it on save, and there are never leftover entries.

Step 4 · Voice & listening

Controls how the agent sounds (text-to-speech) and how it hears the caller (speech-to-text).

Text-to-speech engine (the voice)

Speech-to-text engine (the listening)

Voice ID, your own key & noise suppression

• Voice ID (max 80) — optional specific voice; leave blank for the provider's default (ElevenLabs falls back to its "Rachel" voice).
• Your own voice key (max 500) — optional; same "blank keeps the saved key" behaviour as the AI key.
• Background noise suppression (default on) — strips traffic, fan, and keyboard sounds before transcription, improving accuracy on noisy lines.

Step 5 · Routing & recording

Decides what gets recorded, what happens on voicemail, and when to hand off to a person.

Voicemail / no-answer behaviour

Recordings and transcripts appear in Call Logs exactly as you set them here — the three recording switches are independent, so a call can have a transcript but no audio, or one side of the audio but not the other.

Your own keys vs admin keys

Both the AI key and the voice key are optional and stored securely. The platform decides which key to use at call time in this order:

1. Your workspace's own key. If your plan allows your own keys and your workspace has an active key for that provider, your key is used and usage bills to you.
2. Admin's key. Otherwise the admin's key for that provider (from Admin → AI & API Keys) is used, and usage counts toward your workspace's monthly AI usage limit.
3. Neither. If no key is available, you'll see a clear error — and on a live call, the call is declined so voicemail handles it rather than the caller hearing silence.

In practice: leave both key fields blank and the assistant uses the admin keys automatically. Add a key only when this specific assistant needs a different voice or its own quota.

Managing assistants

From the list you can:

• Pause / Go live — switch an assistant between Live and Paused without opening the wizard.
• Duplicate — copies the full configuration and all skills into a new draft named "(copy)", handy for spinning up a variant.
• Delete — removes the assistant but keeps its call logs readable. Calls already in progress continue; new calls go unanswered.

The list shows per-status counts (all, live, draft, paused) and, for each assistant, its model, voice provider, skill count, and how many calls it handled in the last 24 hours. A "Call logs" shortcut sits next to "New assistant".

Cost cautions

A live voice agent bills three AI meters per call — the AI model (tokens in and out), speech-to-text (seconds of audio transcribed), and text-to-speech (characters spoken) — on top of your WhatsApp/WABA calling minutes. Premium voices and larger models cost noticeably more per minute.

• Keep assistants in Draft until tested. A badly written brief that makes the agent ramble inflates both the AI and voice spend on every call.
• Prefer the lite / mini / haiku models and a concise Pace for high call volume.
• Watch per-call cost on the call detail page, which breaks out the three metered parts: AI usage, transcription, and the spoken voice.
• When you use admin keys, AI usage counts toward your plan's monthly usage limit; hitting the limit blocks further AI calls until you upgrade or switch to your own key.

Troubleshooting