Pick a topic to dive in and start chatting with peers and Glean employees.
Typically, agents will offer to do things they cannot. For example: AI: Would you like me to setup a meeting with the stakeholders? User: Sure AI: I'm setting up a meeting now User: Any luck? AI: Sorry, but I can't actually schedule meetings or access the calendar Well, not with Glean: User: @Glean how is Vercel being used…
Saw that Glean just released a native windows version of the glean app which is real swell; HOWEVER, I noticed it is a .exe which is not great for IT patch management systems like Intune. Will Glean be releasing a .msi/msix version or posting to the microsoft store to help with enterprise management of this app?
I would be so helpful to have modular functionality with agent "steps". So if a step needs to be repeated in multiple places in an agent, it could be created once, and then copy/pasted or inserted into multiple different places. But if an edit is needed, it would only need to be made on the original, and those changes…
Metrics wise, when a user gives feedback on a chat or agent response, we ONLY get reporting on the # of up or down votes, but we don't get to see which prompt it's related to... and we don't get access to the free form text field that the user can fill in when giving that feedback. But, could we? This info would be VERY…
Users generating and sharing content in Glean often need clean, citation-free outputs for presentations, emails, or official documentation. Currently, removing citations requires manual rewriting or follow-up requests to Glean, which slows down frequent content workflows. Implementing a toggle to instantly show or hide…
This is a request for improvement. In short: please consider removing "Access token" from the Miro setup as it seems that it can be retrieved using the other values and currently, providing expired access token results in an unclear setup error. Details: We repeatedly tried to configure a connection to Miro and followed…
I'm building an agent to extract name, company, and title from photos of badges that people took after participating in an event. I've setup a basic agent that (1) reads the file (2) Responds with the 3 info. However, it's not really able to read the files and it breaks on the first step. Would love to have the help from…
Join us on 11/7 at 10:00 AM PT to see AI Outcome Managers (AIOM) Gustav and Nick build Knowledge Management agents that turn scattered content into fast, reliable answers—driving quick wins and everyday productivity gains. What you’ll see: How to connect docs, wikis, tickets, and FAQs so agents deliver grounded,…
We only support metadata for all the Tableau workbooks and other things. But the UI shows we index everything from Tableau. So the request is to provide clear instructions on the above screenshot, as this is confusing users and does not mention that we only support the metadata.
There are many use cases where you may want to use Plan & Execute Steps actions for their access to the agentic engine/superior reasoning relative to Company Search by itself, however, you almost never want to output a response immediately following a company search step. It would be great if, similar to sub-agents, you…
Create an Idea post and tag it Feedback and Ideas so the community can upvote and our product team can prioritize it.
