The idea of sharing an AI session for repeat helpdesk issues sounds useful, especially when the right runbook depends on the environment. What would you want to check before using Ramp Glass for that, particularly when procedures change?
Could Ramp Glass shared workflows help with IT runbooks and handoffs?
Answers
Start with a troubleshooting handoff, not a workflow that changes production. A shared session could give the next tech a record of what was checked, which procedure seemed relevant, and why.
That only helps if the tech still understands the steps. An AI workflow can make it easier to reuse a runbook without knowing what the commands do, so keep production changes behind a separate review.
Reply to this answer
Versioning would make or break this for me. Can someone see which runbook and version a saved workflow relied on? What happens when that runbook is updated or retired? Also check whether edits to a shared workflow are immediately visible to everyone or can be tested before they're released.
With infrastructure code, teams can use branches, tags, and approvals to distinguish tested changes from work in progress. I'd want to know what controls are available here before treating a saved session as a reliable procedure. I haven't seen those product details established.
Reply to this answer
Test permissions before connecting sensitive documents. Use two accounts with different access levels and dummy documents. Have the more privileged account create a session, then share it with the other account and see exactly what appears.
Pay particular attention to information already included in an answer. Blocking the second account from opening the source document may not be enough if the session itself exposes the content.
7a+plaidwrench_685:
I'd also check whether the shared session makes its source references visible. That would help with both the permission test and figuring out when a procedure has gone stale.
Reply to this answer
The interesting bit isn't just finding the runbook. It's capturing why a particular choice was made. A procedure might say to restart one service without explaining why another should be left alone.
That context might be buried in an old ticket or discussion, if it was written down at all. Dates matter too. A temporary workaround can look like a permanent rule if the later decision to retire it isn't included. I'd check what context the workflow actually captures instead of assuming it knows.
Reply to this answer
Get the people who will support it into the trial, even if they aren't the ones choosing the product. They can ask about permissions, updates, and what happens when a shared workflow is wrong. Those details matter more than a slick demo.
2a+fusecrayon_516:
And it's reasonable to decide not to connect an AI tool to the environment if those checks don't give you enough confidence.
Reply to this answer
Add your answer
Answers are reviewed before publication.
6z!socketpulley_427:
That seems like a sensible first test. The handoff context is the part I'm most interested in, rather than having it make changes.