Product changelog and announcements

Today, we are launching two new capabilities with Twilio Conversations– On-demand rule execution for Conversation Intelligence and Conversation metadata for Conversation Orchestrator

On Demand Rule Execution for Conversation Intelligence
On-demand rule execution gives developers direct, programmatic control over when Conversation Intelligence analysis runs. Rather than relying solely on conversation lifecycle events, your application can trigger rule execution at any moment — at agent handoffs, after specific conversation signals, or selectively on conversations that warrant further analysis.

This capability also supports parameter overrides. Operator parameters — such as classification prompts, confidence thresholds, or domain-specific terms — can be overridden at runtime on a per-execution basis, without modifying the stored rule. This means a single rule configuration can dynamically adapt to the context of each conversation — serving a range of scenarios that differ only in one or two parameters, without maintaining separate rules per variation. To learn more, see On-demand rule execution

Conversation metadata for Twilio Conversations
Conversation metadata introduces a native key-value store on the Conversation resource, enabling your application to attach and update structured context alongside the conversation record. Store external identifiers, derived state from one step of a workflow to the next, or per-conversation configuration toggles — all encrypted at rest and returned as part of the standard Conversation resource response, with no separate datastore required. To learn more, see Conversation metadata.

Combining on-demand rule execution with conversation metadata
Used together, on-demand rule execution and conversation metadata enable workflows where context from earlier in a conversation informs what analysis runs next — all within the Twilio Conversations record. A common pattern– your application runs a lightweight rule via the API, stores the result in metadata, then uses that signal to decide whether deeper analysis is warranted — keeping heavier rules conditional on earlier signals  

To learn more about building tiered intelligence workflows using both capabilities together, see Build on-demand workflows with Conversation Intelligence

The Twilio Voice React Native SDK now supports Expo, the framework most new React Native apps start from.

Twilio Studio now includes two widgets for working with Twilio Profiles directly from a flow: Resolve Profile and Create or Update Profile.

Starting September 30, 2026, the Conference list endpoint (GET /2010-04-01/Accounts/{AccountSid}/Conferences.json) will default to returning only in-progress conferences.

Starting October 22, 2026, Twilio Real-Time Transcriptions automatically starts transcription sessions on a second speech provider when a customer’s primary provider doesn’t connect. Failover is enabled for all existing Real-Time Transcriptions accounts on that date, and it routes only within a customer’s existing region. Use of failover is optional and can be opted out of per sub-account in Console.

The Attempt DLR event type will be discontinued after September 22, 2026 at 10:00PM UTC. Learn more about using discontinued Event Types in Discontinued Event Types.  From this point on, Twilio may stop sending new events for the Attempt DLR event types, even if you have a subscription to them. You will continue receiving events normally from any other Event Types in your subscriptions.

Twilio will begin migrating accounts to an updated Voice Recording Storage billing system.

Between Thursday, November 19, 2026 and Monday, January 4 2027, Twilio will enter a Heightened Awareness Period (HAP) to help ensure reliable message delivery during peak traffic times, such as Black Friday and Cyber Monday. 

During this period, message sending rates to U.S. destinations will be monitored and may be limited at the account level for toll-free, short code, and 10DLC numbers.

Accounts that send high volumes of messages from short codes or toll-free numbers may be temporarily migrated to Market Throughput during the HAP. Customers will be notified via email if their account is affected. Market Throughput doesn’t affect 10DLC numbers, which will continue to live on Account Based Throughput.

For more information on how throughput limits may be impacted and what actions to take, customers can reach out to their Twilio account team.

We’re excited to announce that Twilio integration with Salesforce Service Cloud Voice now supports WhatsApp and SMS, alongside existing support for Voice and listed on the AppExchange.

This release comes with the following enhancements:

• Route SMS and WhatsApp conversations through Twilio TaskRouter into the Salesforce Omni-Channel widget, using the same presence, skills, and capacity model as voice
• Handle messaging conversations and voice calls concurrently in a single agent console, with agent presence synchronizing across all channels on work acceptance

To learn more about setup, please see the setup guide.

ElevenLabs voices are now available in Public Beta for <Say>, powered by the Flash v2 and Flash v2.5 models for fast, natural-sounding text-to-speech. Get started by specifying an ElevenLabs voice via the REST API or in a TwiML Bin.