BotID Deep Analysis catches a sophisticated bot network in real-time
By Steven Van ·
Vercel says Deep Analysis spotted ~40 new browser profiles cycling through proxy IPs and blocked them within five minutes.
On October 29 at 9:44am, BotID Deep Analysis flagged an unusual traffic spike on one of Vercel's customer projects: volume rose 500% above baseline, and the traffic looked like it was coming from real human users. Vercel's team had only just started investigating with the customer when Deep Analysis, running on Kasada's machine learning backend, had already identified the pattern and adapted to classify it correctly.
The investigation pointed to a newly spun-up browser bot network. Around 40 to 45 new browser profiles appeared, each with fingerprints and behavioural patterns Deep Analysis hadn't seen before, generating telemetry that looked legitimate. Over the following minutes, the same browser fingerprints began showing up across a range of IP addresses that turned out to be proxy nodes rather than genuine network origins, a pattern real users don't produce. Once Deep Analysis spotted this, it automatically forced those sessions back through verification to collect fresh telemetry, leading to hands-free blocking within five minutes.
Vercel published the incident as a case study for BotID Deep Analysis, its paid detection layer on top of the free Basic check.