Release 1.155.0 | 10th September 2026

✨ Notable features

Voice AI Silence Reminders

• You can now configure a silence interval for Voice AI, after which the bot will speak if the customer hasn't said anything. If needed, add instructions to your prompt so the bot hangs up or transfers after a set period of silence. The bot won't speak if the customer has asked it to wait.

This feature is only available in Speech to Text to Speech mode.

EU AI Act Compliance

• AI usage is now disclosed to the customer when using LLM Chatbot or Talkative Chatbot. This is a requirement of the EU AI Act and applies to all customers.

If you would like to change the wording of your disclosure, reach out to your Talkative CSM. All bots must clearly disclose that they are powered by AI.

More Translation Options

• A new translation provider is now available in Talkative. If your current translations aren't accurate for your use case, reach out to your Talkative CSM to explore switching to our new translation model.

Voice AI Evaluations

• You can now automatically evaluate your Voice AI interactions, making it easy to spot calls that have gone off-piste and act on them. Find it in the Voice AI editor. Every account gets 100 free evaluation minutes, and by default all calls are evaluated until that limit is reached. Reach out to your CSM if you'd like to know more.

Voice AI DTMF Transfers

• You can now configure DTMF-based transfers in Voice AI. For example, transferring a customer to support when they press 1. Since this doesn't rely on AI, the transfer will always happen when the customer presses the button. Configure it under the destinations in the Transfer Tool.

💡 Improvements

• Updated Anthropic Models - the available Anthropic models have been updated to Sonnet 4.6 and Haiku 4.5.

• Interruptible Introductions - introductions in Speech to Text to Speech mode can now be interrupted when interrupts are enabled.

🐞 Bugfixes

• Parking Lot Security - improved security around Parking Lot configuration.