Twilio Changelog

Previously, when a TaskQueue had been configured for multi-reservation of workers, TaskRouter would try to establish an audio connection for every worker that accepted the reservation request, before rescinding all but one reservation. In rare circumstances, that could lead to more than one worker being placed into conference with the customer.

Now, TaskRouter will only establish the audio connection for the first worker accepting the reservation, while all other worker reservations are immediately rescinded.

To learn more about multi-reservation in TaskRouter, refer to our documentation on multi-reservation in TaskQueues.

Build multi-language flows faster and without duplicating widgets by using Liquid template variables in the Language and Voice fields of Say/Play and Gather widgets.

We are officially ending support for the Epic Conference functionality in Programmable Voice. This was a long-running beta program that allowed developers to create a Voice Conference with up to 1000 participants. It did not prove to be as useful to developers as first envisioned and so we are ending support effective today.

Our standard Programmable Voice Conference provides flexible support for multi-party voice use cases with up to 250 participants. Additional details are available here.

The Twilio Debugger now reports on uncaught exceptions or console.error messages that are triggered from the Flex UI or Flex Plugins. In addition to the Debugger, these notifications can also be consumed via the Debugger Webhooks and Alert Triggers.

You can enable the Debugger Integration on the Pre-Release Features Page in the Flex Admin UI. Learn more about how to use the integration by visiting our End User Guide.

This release of Flex UI introduces minor fixes for chat interaction, a new Action for updating agent activity, and UI performance improvements.

For a full changelog, visit the Flex UI release notes.

You can now record voice calls in Flex with two distinct channels, one per party in the conversation. This allows call center managers and quality supervisors to focus on just the customer or just the agent when analyzing calls. The voice player within Flex Insights will separate both channels and highlight who was talking during the conversation.

Visit our documentation for details on enabling and accessing these recordings.

POST request for Twilio <Pay>, once it has successfully tokenized or created a charge, includes the following updates:

• A new parameter called ProfileId - the identifier of the customer object to which the collected payment method is associated. Depending on your connector, this can be used as a token. This value was previously in the PaymentToken parameter if your <Pay> connector is CardConnect, Stripe, or Chase.
• An updated value in the PaymentToken parameter (depending on your connector) - the tokenized value of the payment method.

These changes provide you context and transparency on the value being returned by your selected <Pay> connector when the payment method is tokenized.

If you are already using <Pay>, these changes will not impact you right away. Learn more about the ProfileId, PaymentToken, and their values by <Pay> Connector here.

With the new Play action, you can use pre-recorded audio in addition to text-to-speech to interact with callers. This is especially useful if you have a well-defined IVR script and need to use pre-recorded voices for reasons related to brand and experience. Learn how to use Play in the docs.

The Twilio CLI lets you access all of Twilio’s APIs through the command line to do things like buy and provision phone numbers, send messages and handle incoming calls. It also has a unique plugin model that makes it easy to extend its functionality that we’ve leveraged to build the Autopilot Plugin. You can use it to manage your Autopilot bots, CRUD Tasks, Fields and Webhooks, and add Samples and Field Values directly from a .csv file. Learn how to use the Autopilot Plugin for the Twilio CLI in the docs.

Collect's Prefill feature intelligently skips questions asking for information the user has already provided in the utterance invoking the task. Prefill can be enabled separately on each question by setting the prefill attribute to the Field Name expected. Learn how to use Prefill in the docs.