We are missing AI usage for delivery 2 and 3 in the [ai-usage.md](https://git.ntnu.no/IT2810-H26/T33-Project-2/blob/main/docs/ai-usage.md). I know we used a deploy.md in delivery 2 for backend in vm and trying to identify possible causes why we couldn't set up the backend. Seems like we forgot to document these. If nobody used ai for delivery three then this should be made clear as well. Update the ai-usage documentation for delivery 2 and 3 accordingly.