Founder's Log #2: Have we made a breakthrough?

For the past few weeks, we’ve been experimenting with ML algorithms and mathematical models to see how we can monitor web traffic in real-time and detect potential cyber attacks.

I think we’ve been too focused on trying to use AI as a solution and we overlooked something obvious. There are clear signals we can use to map user behaviours (e.g. IP, sessions etc.). Why not combine these clear signals to reduce noice?

For example, humans are creatures of habit (usually). If Jane usually checks her emails in the morning from her office in London, and one day she suddenly appear to be checking emails in the afternoon from Athens it seems suspicious. We should be able to quickly identify such anomalies by checking a few parameters without the need for ML or mathematical models.