Ada release notes

The Zendesk Support handoff can now be used for email escalations! This enhancement allows human agents to continue escalated conversations in the same email thread as your AI Agent, using the same support email address.

Consult the Zendesk Support handoff guide and the Email configuration docs to learn more.

The homepage dashboard has been redesigned to give you a clearer, more actionable view of your AI Agent's performance, helping you drive containment, resolution, and customer satisfaction.

What's New?​

• Improved Data Visualization – See key performance metrics like Automated Resolution (AR), CSAT, and cost savings in one place.
• Report Previews – Quickly access insights from detailed reports without leaving the homepage.
• Drill-Down Capabilities – Click into reports to explore trends and pinpoint areas for improvement.
• Customizable Filters – Adjust timeframes, channels, and engagement types to focus on what matters most.

Why This Matters​

Previously, tracking AI Agent performance required pulling data from multiple reports. Now, you can see everything at a glance—eliminating the need to piece together fragmented insights.

Rollout​

The rollout of this new dashboard began on March 17th and concluded on March 31st.

The Zendesk Support handoff is a cost-effective way to offer a messaging-like experience for end users and a ticketing experience for agents handling escalated conversations in Zendesk.

Use the new Zendesk Support handoff to integrate with new or existing ticketing flows and provide agents with the conversational context needed to handle inquiries.

Consult the Zendesk Support handoff guide to learn more.

Ada Voice now integrates with more contact center technologies using Session Initiation Protocol (SIP)—reducing telephony costs, improving audio quality, and lowering latency.

• Zendesk Talk: Send calls from Ada to Zendesk Talk using a SIP INVITE. Before transferring, you can create a Zendesk ticket in Ada and pass its ID as part of the SIP transfer—that way agents see the full context as soon as they pick up.
• Gladly: Leverage a SIP INVITE to seamlessly transfer calls from Ada to Gladly.
• Cisco CUBE: Receive calls from a Cisco CUBE using a SIP INVITE, and transfer calls back from Ada using a SIP REFER.

See our updated Help Docs for setup instructions.

We’ve introduced new playback speed and skip features to help you quickly review your Voice AI Agent or bot's call recordings.

What's new?​

• Adjust Playback Speed: Listen at 0.5x, 1x, 1.2x, 1.5x, or 2x speeds.
• Skip in 15-Second Increments: Quickly jump backward or forward to find the most relevant moments.

The Fetch Chat Metadata block can now be used to fetch just the transcript without a summary. In total, the block now has four available outputs:

• Transcript Only — A plaintext transcript of the conversation
• Summary Only — A brief, AI-generated summary of the customer's issue
• Transcript and Summary — A plaintext transcript of the conversation, with the summary appended at the end
• Chatter Info — All data and variables from the conversation.

To learn more:

• For Generative AI Agents: Visit Fetch Chat Metadata block
• For Scripted chatbots: Visit Generate a chat transcript using the Fetch Chat Metadata block

We’re excited to introduce a new way to route phone calls directly from Aircall to your AI Agent. This functionality bypasses the need for a dedicated phone number or SIP domain and reduces extra PSTN call legs—simplifying your Voice setup.

What's changed?​

• Seamless Setup: Route calls from an Aircall Smartflow directly to an Ada AI Agent with minimal set-up required.
• Easy Escalations: Hand calls back to agents on specific Aircall teams when an escalation is required.

To learn more, see our updated Aircall documentation in Integrate your voice AI Agent with your contact center technology.

Transcript summaries now offer more context on the entire conversation. Summaries are also available on their own as a separate output from the Fetch Chat Metadata block.

Transcript summary changes​

Summaries now offer a full-sentence summary of the end user's inquiry, how the AI Agent tried to assist them, and anything else that happened over the course of the conversation. This is available in all handoffs that support sending a summary.

Fetch Chat Metadata block changes​

A new output called "Summary Only" is now available in the Fetch Chat Metadata block for fetching the summary on its own. The original "Chat Transcript" output is now called "Transcript and Summary"; existing implementations using this output are unaffected.

Learn more about the Fetch Chat Metadata block.

The accuracy component of Automated Resolutions (AR) is now measured by incorporating Processes, Actions, Coaching, and Company Descriptions as additional grounding factors.

What's Changed​

Previously, AR evaluation only considered knowledge content to ground the evaluation of accuracy. Now, additional grounding factors are considered.

These enhancements provide a more precise accuracy assessment, reducing cases where conversations were previously judged as inaccurate—particularly when the AI agent leveraged processes, actions, or coaching during a conversation.

How This Impacts AR​

AR should generally stay the same or improve

• More grounding factors now contribute to verify accuracy.
• Conversations that previously lacked knowledge references may now be correctly classified as accurate based on new grounding factors.
• Contradictions between knowledge and other elements (processes, actions, coaching, or company descriptions) may result in a conversation being considered inaccurate and therefore unresolved.

Why We're Confident AR Accuracy Has Improved​

• False positives, where the LLM predicts a conversation was resolved despite the user explicitly saying in the survey that it wasn't, and false negatives, where the LLM predicts a conversation was not resolved despite the user explicitly saying in the survey that it was, have both decreased as a result of this calculation change.

Learn more about how AR is calculated.

We are introducing the Integrations API, which enables developers to connect external applications with Ada and allows Ada users to install integrations directly from their Ada dashboard.

The Integrations API supports the following capabilities:

• Developers can easily integrate external systems, such as Help Centers, with Ada’s suite of Platform APIs
• Developers can publish integrations to Ada’s Integrations Directory, making them available for all Ada users
• Ada users can discover, install, and configure integrations right from the dashboard
• AI Agents responses are augmented through these rich integrations

Whether you’re a partner developer or an Ada user, the Integrations API empowers you to create a more connected and responsive AI Agent. To learn more, see our Integrations API guide.