Bot score false positives: what the logs actually tell you

Bot score false positives: what the logs actually tell you

Bot Management assigns each request a score from 1 to 99. Low means likely automated, high means likely human. Rules typically challenge anything under a threshold, often 30.

Where false positives hide:
— Privacy browsers and extensions. Fingerprint-blocking tools make a real person look scripted.
— Corporate proxies. Traffic re-written by a security appliance loses browser signals.
— In-app browsers. Social media apps open links in embedded webviews with unusual headers.

How to find them: filter events by score range 20–40 and look at the paths. Real people go to product pages, read for a while, then check out. Bots hit the same endpoint repeatedly with no navigation in between.

If a segment looks human, don't lower the global threshold. Write an exception for that pattern, verified by another signal like a valid session cookie. Precision beats leniency.