Box has announced a new “deep scan” functionality that checks files as they are uploaded to identify sophisticated malware and avert attacks.

The new capabilities constitute part of Box Shield, which uses machine learning to prevent data leaks, detect threats, and spot any kind of abnormal behavior. In April last year, Box added a bunch of automated malware detection smarts to the mix, allowing Box Shield customers to spot malicious content that may already have been uploaded to a Box account. However, so far this has leaned heavily on “known” threats from external intelligence databases — moving forward, Box said it will mesh deep learning technology with external threat intelligence to analyze files for malicious scripts, macros, and executables to protect companies from zero-day (unknown) vulnerabilities.

When a user uploads an infected file, Box will quarantine it for inspection, but still allow the user to view a preview of the file and continuing working.

Additionally, Box said it will now provide details of the uploader’s activities, including whether they are an insider or external user, and show a trail of users who have had access to the infected file — this is designed to save security teams time.

Box said that it expects to to make its malware deep scanning feature available to Box Shield customers by the end of the year.