Introduction
Here's an early look at this month's release, starting with Speech Intelligence. A new guided wizard lets you build, test, and publish compliance alerts, and you can now start from Sanas's Global Alerts Library instead of from scratch. Synthetic test datasets give you more control, new RBAC permissions tighten who can do what, and the Agent Portal dashboard now matches the supervisor view, just to name a few.
Updates for more products are on the way, so check back soon!
What’s New
Speech Intelligence
Sentiment Faces: Now color-coded in Live Tracking.
Transcript Transfer to Azure: Send transcripts to Azure Blob Storage via Entra federated credentials. No secrets stored.
Desktop App Configuration: Choose per group which post-call sections show in the Desktop App and Agent Web Portal.
Feature Management Changes: Toggle each Speech Intelligence capability (summary, action items, discussion points, compliance alerts, sentiment, S3 transfer) from its own section.
Create Alert Wizard: Compliance Alerts now has a guided wizard to build, test, and publish alerts, with Regex vs. LLM result comparison.
Sanas Global Alerts Library: Tenants can now copy Sanas-authored alerts from a shared library, customize them, and promote them to Production.
Test Datasets Synthetic Wizard: Supports audio seeds, manual tagging, and finer pattern-injection controls
RBAC Configurations & Surfaces: Alert, transcript, and Speech Intelligence configuration actions now have their own dedicated permissions.
Agent Portal Dashboard: Now matches the supervisor view for metrics and alert navigation.
Enhanced Sentiment Faces for Speech Intelligence
Live Tracking’s sentiment faces (good, neutral, bad) are now color-coded, making it easier to read sentiment at a glance.
Transcript Transfer to Azure for Speech Intelligence
Transcript Transfer now supports Azure Blob Storage as a destination. It authenticates through your Entra app with federated credentials, so no client secrets or SAS tokens are stored in the Portal.
You can test the connection before saving, and built-in guidance helps you configure the Azure trust settings (issuer, subject, and audience). Existing S3 destinations aren’t affected.
Desktop App Configuration for Speech Intelligence
You can now choose, per group, which post-call sections appear in the Sanas Desktop App and which appear in the Agent Web Portal. Hidden sections are still generated and stored. They just aren’t displayed.
Feature Management Changes for Speech Intelligence
Speech Intelligence capabilities (si_* keys) now have their own section in Feature Management. You can turn each one on or off independently: summary, action items, discussion points, compliance alerts, sentiment, and S3 transfer.
Note: Speech Intelligence must be enabled first for any of these to work.
Create Alert Wizard for Speech Intelligence
Compliance Alerts now has a guided create-alert wizard.
Build an alert in four steps: describe it, generate the detection, test it on a chosen Test Dataset, and publish. Regex and LLM results are shown as separate cards so you can compare them, and a limited test run comes before Production.
Sanas Global Alerts Library for Speech Intelligence
Centrally authored alerts are now shared across tenants.
For Sanas admins: Create and manage alerts in the new Tools catalog.
For tenants: Browse them in the Sanas Library under Compliance Alerts, copy one into your group library, edit it in the wizard, and promote it from Non-production to Production.
New update!
The Global Alert Library now lets you reopen drafts and edit a published alert by saving it as a new version. Each alert also shows its fork count. This release also improves the dataset picker and fixes issues in the matrix and adoption views.
Test Datasets Synthetic Wizard for Speech Intelligence
The synthetic wizard now accepts audio uploads as seeds and adds manual tagging, final seed review, full-dataset paging, and finer pattern-injection controls (placement, per-pattern targets, ground-truth category tags).
Fixed word limits, counts, and description requirements based on staging feedback.
New update!
The Dataset library now shows the creator’s display name; generation shortfall is visible from production feedback.
RBAC Configurations and Surfaces for Speech Intelligence
Alert, transcript, and Speech Intelligence configuration actions, including Reset to parent, now each have their own permissions instead of falling under the generic group-settings update permission.
Note: RBAC is short for Role-Based Access Control.
Agent Portal Dashboard for Speech Intelligence
Dashboard metrics and alert navigation now match the supervisor experience, with the same sort order and alert-only buckets shown when the call overview is empty.
Support
Need help? Get in touch with our Support Team for assistance.