Users can now subscribe to automated KBo analysis for any saved search in Conversation Explorer. KBo will continuously monitor new matching conversations and deliver synthesized insights straight to your inbox—helping you stay ahead of emerging trends, risks, and opportunities.

KBo will analyze new conversations (minimum of 1, up to 50) and generate a structured summary based on your input.

Each email includes a link to open the full analysis in KBo within the platform.

KBo analysis excludes all PHI/PII. We have taken extra care to make sure no sensitive data makes it into emails.

Users can now rename montages on custom dashboards! This allows users to name the montage to fit the purpose of the dashboard. Renaming the montage on the custom dashboard will only change the display name on that dashboard, but will not change the montage name anywhere else in the platform.

Users can now rename insights cards on custom dashboards. This is especially helpful now that it is possible to add filters to insights cards, because users can name them appropriately to reflect the data that is visible.

We are absolutely delighted to announce a massive upgrade to the reporting power of Signals: chart-level filtering! Imagine being able to compare, on a single custom dashboard:

Users can now see any of the above (and more!) with chart filters in Signals.

How it works:

Note: Users can apply chart filters on standard dashboards (e.g., Eddy Reports, Themes & Topics, etc.), but must pin the report to a custom dashboard for chart filters to persist across sessions.

For more details on How it Works, see our Help Center article on using chart filters.

As mentioned in our last release notes, we started running the Outline model on audio conversations on 2/28.

Here’s a refresher on the model:

The Conversation Outline Model divides each audio conversation into distinct, labeled “chapters” that represent common stages or moments within a customer interaction. By identifying these consistent conversation segments, the model creates a structured outline that supports reporting, filtering, and navigation across transcripts.

This segmentation enables users to:

The result is a clear, repeatable structure that transforms raw transcripts into organized, searchable data for analysis, coaching, and operational insight.

Note: Conversation Outline is only available on audio conversations (not chats) at this time.

Now, you can use the Conversation Outline Model in two ways within the platform. Read on for details!

Users can now use the Outline Model in Conversation Details to quickly navigate to specific parts of the call!

In the Transcription tab, there is now a “Jump to” dropdown. Click any of the segments in the dropdown to automatically scroll to that part of the conversation. These segments are also clearly defined within the transcript itself.

Users can now filter to calls that contain or do not contain specific Outline segments, using the Outline filter. For example, users may want to find calls where patients called with a specific problem, but no solution was provided on the call. Simply use the Outline filter to find calls that contain a Problem statement and do not contain a Solution Provided.

Note: Outline Filter is not yet available in chart filtering, but will be available in a future release.

We previously introduced the ability to group by agents on standard model results reports. This is especially powerful for assessing individual agent performance for each standard quality model. Now, we’ve added that same group by agents capability onto our automatically created model results reports. Simply go to the report menu, select “Configure”, and group by agents to see how model results compare across agents.

Note: Users can group by agents on the model results reports on standard dashboards, but must pin the report to a custom dashboard for grouping selections to persist across sessions.

You can now leave comments directly on conversations, tag teammates to notify them, and reply to existing comments, making collaboration easier than ever.

How it works:

As always, existing permissions apply. Only users with access to a conversation can view and comment on it.

Previously, when users clicked into the platform from a “Subscribe to Search” email, the saved filter (including its original date range) was applied. However, subscription emails report the total number of new conversations from the past day, week, or month — often a different timeframe than the saved filter’s date range. This could result in a mismatch between the conversation count shown in the email and what users saw in the platform.

For any emails received after Friday, March 6th, when users click into the platform from a subscription email, the same date range used in the email will automatically be applied. This ensures the conversation count matches between the email and the platform.

Users will see a toast message confirming this update: “Filters from your subscription email have been applied to this view.” Again, this will only apply to emails moving forward!

We’ve added Client Call ID to the options users can select when exporting Model Results graphs to CSV!