In all generative AI Agents and scripted bots, you can now configure two new settings in your CSAT surveys:

• Optional 10-point numeric scale: You can now configure surveys about both AI Agents and human agents to be on a 10-point scale, in addition to the existing 5-point and binary scales.

• Non-Ada Glass handoff logic: Previously, when customers got handed off using a handoff method that didn't use an Ada Glass integration, bots would assume the conversation was ended and show customers a CSAT survey.

  We've improved the logic in this scenario, so customers will no longer see a survey before the end of their conversation.

Additionally, after releasing some CSAT survey improvements to generative AI Agents earlier this month, we're happy to release those same updates to scripted bots too.

We've made writing content easier by merging Text and Shuffle blocks. Now, within your Text blocks, you can generate message variations using AI, or add them manually.

We have released test scenarios, which you can use to validate that your AI Agent is providing the correct content to your different audiences (e.g., to provide different greetings to customers based on their region). You can set variable values in the test chat, and chat with your AI Agent as if you were a user with those variables set.

You can now safely authenticate your HTTP requests with static tokens without exposing the value in your Ada dashboard. Navigate to the Token vault page in your dashboard settings to add new tokens.

For more information, see Securely store API tokens using the Token Vault.

The Scheduled block and Schedules setting are now available across all Ada dashboards! You can now build content to provide relevant customer experiences around your organization's operating hours.

For more information, see Manage your AI Agent's basic dialog

Your bot now automatically categorizes conversations with similar customer inquiries and creates topics to categorize them into, which you can then manage as required.

For more information, see our Help Docs for generative AI Agents and scripted bots.

Got a lot of articles on your Knowledge page? You can now search by article name, or apply filters to find articles more easily.

For more information, see our Help Docs for generative AI Agents and scripted bots.

We've added the ability to search for individual blocks in the Action Apps section of your block drawer. Previously, you could only search for all Action App blocks at once by searching for <block:action_integration>, but now you can search for specific Action App blocks (e.g., you can find all of your Answers that contain the Answer Utilities block by searching for <block:action_integration:answer-utilities>).

For a list of block names you can search by, see Search by Answer block.

If you're transitioning from using one bot to another, you can now roll out your second bot progressively, specifying what percentage of customers you'd like to see each of your bots, or no bot at all.

For more information, see Control how many users see your bot.

In addition to Google's speaking voices, you can now choose one of OpenAI's voices for your bot to use with your callers.

For information, including differences between how Google and OpenAI voices behave, see Choose a speaking voice for your bot.