Twilio Changelog

Voice Insights – Intelligent Discovery AI Assistant (IDA), available on the Call Insights page in Twilio Console, lets you explore your voice data simply by asking questions in natural language. The Assistant helps you troubleshoot voice-related issues, including deliverability (connection rate), answer rate, networking related issues and other behavioral issues such as silent calls. It can assist with data analysis and dive deeper into your voice data to find potential root causes. The Assistant provides recommendations to solve voice errors and potential issues. You can provide feedback on each response and the overall Assistant experience. Activate Voice Insights Advanced Features to use this feature.

For more information, please refer to the Voice Insights reference docs. 

As of May 21, 2025 the Legacy Boku Identity APIs Infrastructure in the United States will be migrated.

We would like to inform you about upcoming changes to Legacy Boku Identity API’s production infrastructure in the United States. These changes will not impact your current service. Our goal is to minimize impact to your current integration to our services. These updates are part of our ongoing commitment to improving performance, scalability, and reliability for all our customers.

If you have whitelisted or hardcorded our IP addresses instead of URLs, please make a note of our new IP addresses. These new IP addresses will take effect once we perform the switchover to AWS cloud services.

Our support team is available to assist you with any questions or concerns at accsec-onboarding@twilio.com.

Thank you for choosing Twilio. We appreciate your partnership and look forward to continuing to provide you with exceptional service.

What’s changing:
Programmable Chat in Flex is reaching end of life (EOL) on June 1, 2026. After that, it may stop working as expected and won’t receive any more enhancements, bug fixes, or security updates. It's had a good run—but it’s time to pass the baton.

Customers still using Programmable Chat in Flex will need to migrate to Flex Conversations, Twilio’s supported messaging platform for channels like Webchat, SMS, and WhatsApp. With key feature gaps now closed, like the GA of Webchat 3.x, there’s no reason not to make the switch.

To support this transition, the following resources are available:

• Flex Conversations Overview

• Migration Guide from Programmable Chat to Conversations

No surprises here—both Programmable Chat and Conversations use the same billing model.

For any questions or support during the transition, customers should reach out to their Twilio account team.

<Connect><ConversationRelay> is a TwiML noun and Service Provider Interface (SPI) for building genAI-powered virtual agents. ConversationRelay makes voice AI integrations straightforward, so you can focus on delivering the self-service experiences customers expect without diving into the complexities of voice technology. It combines low-latency speech-to-text and text-to-speech capabilities with human-like conversational pacing, orchestration, and seamless interruption handling through a websocket API. Bring your own large language model (LLM) and let Twilio take care of the rest. 

New features included with this release:

• ElevenLabs is now the default TTS provider. A new TwiML param “elevenlabsTextNormalization” specifies whether or not to apply text normalization while using the ElevenLabs TTS provider.

• New TwiML param “reportInputDuringAgentSpeech” specifies whether your application receives prompts and DTMF events while the agent is speaking

• New TwiML param “debug“. Options are debugging, speaker-events, and tokens-played. The debugging option provides general debugging information. speaker-events will notify your application about agentSpeaking and clientSpeaking events. tokens-played will provide messages about what's just been played over TTS.

Check out the blog post to learn more. If you’re ready to dive right into code you can find the docs here. We can’t wait to see what you build!

Voice Intelligence has been renamed to Conversational Intelligence – reflecting our expansion into analyzing conversations across all customer touchpoints. Businesses can now unlock unified AI understanding across Voice, Messaging & Virtual Agents. 

Building on the foundation of Voice Intelligence's AI-powered transcription and language analysis capabilities, Conversational Intelligence introduces powerful new features designed for today’s cross-channel engagement landscape:

🤖 AI Agent Observability with ConversationRelay (Generally Available)

A native integration with ConversationRelay empowers you to monitor and refine interactions between customers and Twilio-powered AI agents to ensure quality and fine-tune performance over time.

✨ Generative Custom Operators (Public Beta)

LLM-powered custom operators let you easily define use cases like quality assurance checks, compliance flagging, or competitive insights — tailored to your business — all with just a few simple instructions. 

💬 Conversational Intelligence for Messaging (Private Beta)

Gain full visibility into customer interactions across SMS, WhatsApp, and web chat. Analyze messaging conversations alongside voice data to uncover trends, preferences, and patterns throughout the entire customer lifecycle.

Want to learn more?

• 👉 Explore the Conversational Intelligence docs
• 👉 Get building using Conversational Intelligence with ConversationRelay
• 👉 Create your first Generative Custom Operators
• 👉 Request access to the Conversational Intelligence for Messaging Private Beta

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Important Note: Voice Intelligence customers do not need to take any action as a result of these changes, and Transcription & Language Operators will continue to function normally. 

Twilio, by default, handles standard English-language reply messages such as STOP, UNSUBSCRIBE, END, QUIT, STOPALL or CANCEL for US and non US Long Code numbers, in accordance with industry standards. As of May 13, 2025, Twilio added REVOKE, and OPTOUT as additional opt-out keywords.

Learn More: Update to FCC's SMS Opt Out Keywords

Beginning today, agents can now leave and pause a task for any Flex Conversations channel, such as SMS or webchat. We’re expanding this functionality from only email tasks to all Flex Conversations tasks. 

Agents can also enter wrap-up after they leave or pause a task, so they can perform final wrap-up activities before completing a task. This ensures that agents have time to document their actions and also provides a more accurate report of their status.   

In addition, administrators and developers can now configure a webhook URL and subscribe to notifications about Interactions API events that occur during conversations. 

Event notifications can provide information about customer interactions or enable developers to trigger additional workflows after certain events. For example, when used with leave and pause functionality, Interactions event notifications enable you to trigger custom logic when a customer responds to a conversation the agent has left.

With SAML-based SSO, users can now automatically join a primary team when they sign in to Flex. In addition, when supervisors sign in to Flex, they can automatically gain oversight of multiple teams as the assigned owner. By using attributes from your identity provider (like Okta), this update simplifies team assignments right from the start. 

To learn more about Flex Teams and how to set up SSO, see Teams in Flex and Configuring SSO and IdP in Flex.