Has anyone used or explored how can we use ROVO for Dynamic License Management

[sshalabh]

We want to manage license on rotation but we want the AI to choose the person with least posibility of returning in next 15 days to the product

[Tomislav Tobijas · Accepted solution]

Hi @sshalabh,

If I understood this correctly, you'd like for Rovo/AI to add or remove users (licenses) based on some kind of analytics and prediction.

I haven't tried using Rovo to at least get me that information, but from what I've seen so far, it's probably the case that it currently isn't able to do that. 👀

I do know that automation in admin.atlassian.com is in the works, which will help automate some kind of user management tasks, but AI isn't included in this part (at least not yet).

However, with all the development that's going on around AI and Rovo, it might be that something like that will be available relatively soon, but that remains to be seen.

Anyway, for the first step, I'd recommend testing on existing data will Rovo at least provide results related to user activities and some kind of recommendation (relatively unlikely). Building on that, maybe some kind of API calls could be added to then remove this user, but again, I haven't tested this myself.

I'll move this to the Rovo forum group and then maybe someone from there could chime in.

Cheers,
Tobi

[Dr Valeri Colon _Connect Centric_]

Welcome to the community @sshalabh I'll add to what Tobi is say, Rovo can’t manage licenses or predict user return likelihood on its own. It has no access to login-frequency data, product-usage logs, or license assignments. If you export that data yourself (via Atlassian Admin APIs), you can feed a processed list into a Rovo agent and have it rank candidates, but the decision logic must come from your data pipeline. True dynamic license rotation requires external scripts or Automation—not Rovo alone.

[Ed Letifov _TechTime - New Zealand_]

@sshalabh 

Disclosure: I represent a vendor of User Management Marketplace app that does user and license management, with scheduled actions giving you ability to take the unused licenses back.

Indeed, at some point AI could be unleashed on this area... However what you do need to consider is who is going to take responsibility if the AI un-licenses someone uber-important (think "your CEO", but really, think "the last remaining deployment engineer" or something similar).

Given the widely accepted possibility of AI hallucinating, whoever writes the prompt should not only be a true believer, but also a heck of salesman to convince their manager to accept the responsibility for failure.