November 3, 2026
Release version: 567
This release adds Otter.ai and 15Five connectors, introduces a content trigger for new artifact versions, brings auto mode routing policies and GPT-6.1 Terra to Glean Assistant, and makes Deep Research available as a skill. It also strengthens access checks for triggered deliveries and ships reliability fixes across chat, agents, and connectors.
Major launches
Otter.ai connector now in public beta
The Otter.ai connector is now available in public beta. Connect Otter.ai to make meeting knowledge searchable in Glean — including transcripts, summaries, action items, insights, outlines, participants, and channel context — and available to Glean Assistant.
How to access: Admins can set up the Otter.ai connector from the Admin console.
15Five connector now available to all customers
The 15Five connector is now available to all customers. It indexes public High Fives and weekly Priorities from 15Five, so recognition and goals are searchable alongside the rest of your company knowledge.
How to access: Admins can set up the 15Five connector from the Admin console.
Trigger agents when a new artifact version is published
Agents can now run automatically when a new major version of a Glean artifact is published, so downstream workflows stay in sync with the content your team produces.
How to access: In the agent builder, choose the New artifact version content trigger and pick the artifact to watch.
Agents can access ServiceNow with a service credential
Admins can now create a ServiceNow service credential, so agents can access ServiceNow as a shared service account instead of relying on each user's individual authorization.
How to access: Admins create the ServiceNow service credential from the Admin console.
Deep Research is now a skill
Deep Research now runs as a skill, so you can start thorough, report-style research the same way you use other skills.
How to access: Start it from the skill picker or type /deep-research in the composer. Available to users with access to skills.
Review everything in a chat from the new context panel
Glean Assistant now has a side panel that shows everything used in the current chat in one place — the agents that ran, the outputs created, and the sources cited — so you can quickly review or revisit sources and outputs from the chat session.
How to access: This feature is in managed rollout. Admins can enable it from Feature rollouts in the Admin console ahead of general availability planned for November 24, 2026.
Unsubscribe from Glean emails without signing in
Recipients can now unsubscribe from eligible Glean email categories without signing in, including after losing access to their Glean account. Unsubscribe links require confirmation and apply only to the linked notification category. Required account and security emails are unchanged.
How to access: No action needed. Use the unsubscribe link in eligible Glean emails.
AI & model updates
GPT-6.1 Terra now available
GPT-6.1 Terra is now available as a live model for supported Glean experiences, giving your team another frontier model option.
How to access: Select GPT-6.1 Terra from the model picker where supported. Admins control model availability in the Admin console.
Auto mode picks the right model for each request
Auto mode in Glean Assistant now picks the model and reasoning level for each request, so users no longer need to choose a model themselves. Admins can assign an Economy, Balanced, or Frontier routing policy to the whole organization or to individual departments. Admins can also turn on automatic cost optimization, which moves Frontier users to Balanced at 75% of their monthly usage limit and moves users to Economy at 90%. Responses show which models were used, and intermediate steps note when Assistant switches models. Because auto mode now sets reasoning effort automatically, the Fast, Adaptive, and Thinking modes have been removed; users who pick a specific model can instead set its reasoning effort from Low to Max, where supported.
How to access: Available on deployments that use Glean-hosted model keys. Admins assign routing policies from the Admin console.
Critical updates
Slack-triggered deliveries check channel access
Slack-triggered webhook and agent channel deliveries now check that the subscriber can see the Slack channel before delivering a message, keeping notifications aligned with Slack permissions.
How to access: No action needed. The check applies automatically.
Zendesk-triggered deliveries require verified access
Zendesk-triggered agents and webhooks are no longer delivered when a Zendesk webhook payload can't be checked for the subscriber's access.
How to access: No action needed. The safeguard applies automatically.
Enhancements
Let Slack agents respond to mentions in any channel
Admins can let a Slack agent respond to mentions in any channel. By default, an agent responds only in channels it has been added to and in direct messages.
How to access: Admins enable this in the agent's Slack settings.
Rename chats from the sidebar
You can now rename a chat directly in the sidebar: double-click its title, or choose Rename from its menu, then type a new name and press Enter.
How to access: No action needed. Available in the chat sidebar.
Fewer repeated follow-up offers
Glean Assistant no longer repeats the same follow-up offer within a chat, such as asking who to share a document with after every edit.
How to access: No action needed. The improvement applies automatically.
Independent agents can use user-authorized tools
Independent agents can now use tools that require user authorization. When an agent needs such a tool, an approver can lend their access to that call through an approval request.
How to access: No setup needed. When an independent agent needs a user-authorized tool, the approver receives an approval request.
One-click microphone access in Chrome
In Chrome, users whose browser blocks the microphone for Glean now see a prompt on the voice button — at most monthly — that lets them allow the microphone in one click, without opening browser settings.
How to access: No action needed. The prompt appears when the microphone is blocked.
Refreshed side panel header in the browser extension
The browser extension's native side panel has a refreshed header with a new three-dot menu for chat history and settings, plus a background highlight on the chat, search, and agents selector.
How to access: No action needed. The update applies automatically.
Edit blocked messages to continue the conversation
When a chat message is blocked by your organization's content policies, new messages are disabled until you edit the blocked message, and the error now tells you to edit your message to continue the conversation.
How to access: No action needed. Applies when your organization's content policies block a message.
Tidier tool connection cards in chat
Once you connect or skip a tool mid-response, its connection card shrinks to a single line — "Connected to X" or "Continued without X" — so it no longer interrupts the view of what Assistant is working on.
How to access: No action needed. The change applies automatically.
Legacy chat route redirects to the new home page
Users with the new home page who land on the legacy /chat route are now redirected to the new home page.
How to access: No action needed. Applies to users with the new home page.
Citations point to the exact page or slide
Chat citations for PDFs and slide decks can now point to the specific page or slide, so you can jump straight to the cited content.
How to access: No action needed. The improvement applies automatically.
Add a personal message when sharing agents
When you share an agent and choose to notify teammates, you can now add an optional personal message. The message is included in the email and Slack notifications sent to the people you share the agent with.
How to access: Add the message in the share dialog when notifying teammates.
See which models are Standard on credits-based plans
For customers on credits-based plans, the Admin console now clearly shows which models are Standard.
How to access: Admins can view model designations in the Admin console.
Use your own Salesforce External Client Apps
Salesforce connections can now use customer-owned External Client Apps while keeping user authorization scoped to the connector and app.
How to access: Admins configure this in the Salesforce connector setup.
Limit Power BI indexing to specific workspaces
Admins can now limit Power BI indexing to specific workspaces by entering workspace IDs in the Power BI connector setup. Leaving the field empty keeps indexing all workspaces.
How to access: Enter workspace IDs in the Power BI connector setup in the Admin console.
Map Trello users through Trello Enterprise
Trello admins can now map users through Trello Enterprise member emails instead of Jira or Confluence Cloud, on new or existing Trello instances.
How to access: Choose the Trello Enterprise option in the Trello connector's user mapping settings.
Faster, more reliable bulk agent deletion
Admin bulk deletion of agents is faster and more reliable for large comma-separated values (CSV) files. Agents are deleted in batches, and one failing agent no longer blocks the rest.
How to access: No action needed. Applies to bulk deletion in the Admin console.
Workflow-run events note uploaded files
Workflow-run customer events now indicate whether the initiating user message included uploaded files, improving audit visibility.
How to access: No action needed. The field appears in workflow-run events.
Spark pipelines pull images from your deployment's Artifact Registry
Spark pipelines on Google Cloud Dataproc can now pull their runtime images from the deployment's own Artifact Registry repository without manual Identity and Access Management (IAM) grants.
How to access: No action needed. Applies automatically to Google Cloud Platform (GCP) deployments.
Fixes
Chats no longer stay busy after a response finishes
Fixed an issue where Glean Assistant and agent chats could stay busy with the send button disabled after a response had already finished, until the page was refreshed.
Agents available for users with model overrides
Fixed an issue where agents were unavailable for users whose assistant model is set through a model override.
Jira Service Management requests use your form selections
Fixed an issue where Jira Service Management request creation could apply AI-generated field values instead of the values selected in the request form.
More reliable OAuth connections after inactivity
Fixed custom actions that use the OAuth client credentials grant failing with authentication errors after a period of inactivity when the provider doesn't return a token lifetime (for example, Salesforce). Glean now refreshes these tokens automatically and more often.
Connectors no longer stuck in initial sync
Fixed connectors (for example, Jira Data Center) staying in "Initial sync in progress" after the initial crawl had completed.