If the alerts shows he accessed 37 resources, then most likely he did...
He might only remembers 2/ or only directly accessed those 2, but actions under his account accessed more...
What are those other resources? do you have logs on those specific resources or some of them which he claims he did not access to see if you can find access evidence there as well?
Did he use any new apps that might have accessed those resources under his account without him knowing it?
Any chance of malware on this device?
Excluding his account for this detector will prevent this detector from reporting this user ever again, doesn't matter from where or to where.
I would not do that until I have proper explanation of what happened and if it is legit or not.