Knowledge base
Checks on every message
Every message you receive scored for spam, phishing and AI writing, with each check behind the scores listed.
Details
- The Details tab of a message holds its checks: six tiles for Spam score, Phishing score, AI score, Trackers, Sender and Encryption. The worst finding is selected when it opens, and What we look for, under a tile, lists every check behind it as flagged, seen, clear, not seen, not run or no result. Why we flagged this, on a warning, opens the same place.
- Spam, phishing and AI writing are scored out of 100. Spam reads as unlikely under 35 and likely from 60, adding up signs such as a failed sender check, a mailing list with no way to unsubscribe, a Re: that answers nothing, a subject in capitals, money bait and link shorteners. Phishing reads as worth a second look from 30 and a scam from 60. AI writing reads as possible from 35 and likely from 60, and under 40 words is too short to tell.
- The scores describe a message; they do not move it. Four things alone file mail in Spam: a virus, the spam filter’s verdict, a DMARC failure on a domain that enforces it, and a spam header from the server that handed the message over.
- Trackers counts the open pixels and tracked links in a message, names the hosts they report to, and says whether a pixel could see you open it, which depends on whether remote images are showing.
- Mail that arrived before the checks existed, and mail you imported, is scored as it is read, without a model and without the headers, so the rows that need headers say No result. Mail you sent has no spam, AI or tracker score.