Handoff reasons in the Data Export API and MCP
The AI Agent now records why it hands a conversation to a human agent. It records a reason_code from a fixed list of ten values, or null when it records no code. It also records a reasoning text when one is available.
The v2 Conversations endpoint of the Data Export API now returns a handoffs array on each conversation. Each entry is a handoff attempt with its outcome_status, reason_code and reasoning. The get_conversation MCP tool returns reason_code and reasoning.text on its handoff entries. The new HANDOFFREASON filter in get_available_filters filters conversations by reason code.
Use the reason codes to find handoffs that you can reduce, such as a failing Tool or a gap in your Knowledge.
See Handoff reasons for the list of codes, and the Conversation object reference for the full schema.
AI Agent reads receipts that show a masked card number
Your AI Agent now reads receipts, invoices, and order confirmations that show only the last 4 digits of a card. Before this change, the Agent withheld these images as payment cards.
What’s changed
The rule that withholds payment cards now tells a receipt apart from a card.
- Masked card numbers are read: a receipt or order confirmation that names the card only by its brand or its last 4 digits is described like any other image.
- Payment cards are still withheld: an image of a payment card is withheld, including a saved card on screen with a masked number.
- Full card numbers are still withheld: an image that shows a full card number anywhere is withheld. See Images.
Opening line for Voice Proactive Outreach
A Voice Proactive Outreach that runs a Playbook can now start every call with an Opening line: the exact words that the end user hears the moment they answer.
The line does not depend on how the Playbook begins, and it plays before the Playbook’s first step is ready. Write it once in the default language of your AI Agent. Include your business’s registered name, along with any AI disclosures or recording notices required by law. Every call then opens with those exact words before the Playbook runs.
When you save, the Agent translates the line into the other languages that it speaks on Voice, and you can modify each translation. The line is the first message in the conversation transcript, so the record shows what the end user heard.
The end user also hears the Agent sooner. The line plays as soon as they say hello, or after 2 seconds of silence, while the Playbook’s first step is still being prepared. Answering machine detection no longer holds the call before the Agent speaks. A call that reaches a voicemail system still ends without a conversation.
See Opening line on the Voice Proactive Outreach page.
Find and fix Playbook errors from the page header
The Playbook editor now lists every error in one menu in the page header. AI Managers can go straight to each error and fix it before they activate the Playbook.
What’s new
The N errors to resolve menu sits next to the Active toggle and replaces the error count on the canvas.
- Grouped list. Errors on the Playbook’s own fields, errors on steps and sections, and warnings show in separate groups.
- Go to the error. Select a row to scroll to the step, section, or field and place the cursor in it.
- Errors on the field. The When to use this Playbook field shows its error below the field.
Changed checks
The editor now shows these problems as errors on the canvas. The API already rejects them when a Playbook is activated.
- An
IForELSE IFbranch with no steps. Before, the editor showed a warning. - A
GO TOstep with no target. - A section with no steps.
An inactive Playbook with these errors still saves. Fix them to activate the Playbook.
See Find and fix errors.
Voice: set how long your AI Agent waits after a key press
A new Keypad entry timeout (seconds) setting is on the Voice Configuration tab. It controls how long your AI Agent waits after the last key press before it uses the digits. Set it between 2 and 7 seconds. Pressing # submits the digits at once.
The timeout applies wherever an end user enters a value on the phone keypad. That includes Ask steps in Playbooks, Action inputs, and keypad capture in Greetings, Handoffs, and Processes. A shorter timeout makes keypad entry feel faster. A longer timeout gives end users more time between keys.
While an end user is entering digits, the Agent does not play the caller-silence re-prompt. This holds even when Maximum caller silence (seconds) is shorter than the keypad entry timeout.
The setting starts at 5 seconds for AI Agents that already use Voice, so nothing changes until you edit it. See Configure how long your AI Agent waits after a key press.
Set the message the AI Agent says when it withholds an image
You can now replace the sentence your AI Agent says when it withholds a sensitive image. Type your sentence in Message when an image is withheld under Image redactions. The Agent translates your sentence when the conversation is in another language. It uses your sentence for identity documents and other credentials when it can offer a human agent, and the default sentence otherwise. Leave the field empty to keep the default reply.
For more information, see Redaction settings.