Artificial intelligence has been discussed in cyber security for years, largely through the lens of incremental improvement. Better detection, faster analysis, more automation in repetitive tasks, etc. What has not fundamentally changed, not to this level, at least, is the tempo of the threat landscape itself… until now. The launching of Mythos/Fable (and other Frontier AI) represents a clear changing point.

The UK rail sector faces complex cyber security challenges under the NCSC Cyber Assessment Framework (CAF). This includes protecting operational technology (OT), managing legacy signalling systems, and securing a highly fragmented supply chain across Network Rail, TOCs, and ROSCOs. The Department for Transport applies CAF within the NIS Regulations, prioritising passenger safety alongside cyber resilience, particularly during CP7 infrastructure programmes.

I've had this conversation more times than I can count. A client comes to us having already deployed SD-WAN across their estate, and now they want to layer on cloud security. What should be straightforward always turns into a rework project. Here's how I think about it — and how to avoid the same trap.