UK start-up Cyberlytic has announced what it calls a unique offer to small and medium sized enterprises.

The company is making available to SMEs its new ‘Cyber Threat Profiler’, originally developed for the UK Ministry of Defence (MoD), which it says “uses artificial intelligence (AI) techniques in a way that is both faster and more sophisticated than ever before”.

Until now, Cyberlytic claims, signature-based detection has proved insufficient to protect against Zero Day and polymorphic attacks, but “through AI and machine learning technologies, the ‘Cyber Threat Profiler’ detects and prioritises attacks based on the risk they pose to a company’s IP and data, and can therefore, detect attacks that may have been overlooked by other systems”.

According to its creators, the Cyber Threat Profiler implements both anomaly detection and a patented classification approach to detect and effectively triage cyber-attacks, whereas other security tech firms are only using AI to perform anomaly detection.

Stuart Laidlaw, co-founder and CEO of Cyberlytic, adds: “We created this solution to help security teams identify and evaluate the risks posed by advanced cyber-attacks. Based on research we completed for the UK MoD, we’ve developed an effective way of using AI to detect and prioritise web attacks. Our data science approach evaluates the risk posed by an attack by looking at the sophistication and the capability of the attacker. Importantly, our software also assesses the effectiveness of each attack, combining these characteristics to create a risk score. This is achieved through real-time deep packet inspection of all HTTP traffic – something that no-one else out there is doing.”

Laidlaw’s background includes experience working for Fujitsu, BAE Systems and government agencies. Cyberlytic’s co-founder and CTO, St John Harold, has also worked with several FTSE 100 companies as a security consultant and more recently, has provided expert security advice to UK government departments and agencies.

They argue that companies can no longer rely on human intervention alone to quickly and efficiently detect threats. They cite estimates that more than 20,000 compromised websites are identified every week, and that the average security team takes around 220 days to identify and respond to a successful cyber attack, whereas the ‘Cyber Threat Profiler’ takes response time down to a matter of seconds.