Vectra AI launches Vectra AI Pro to feed AI agents better attack signals

Threat detection firm Vectra AI has introduced Vectra AI Pro, a solution designed to enhance artificial intelligence agents utilized within security operations centers (SOCs) with more dependable information about attacker behavior. As security teams have integrated AI into their workflows over the past year to keep up with rapid attacks, agents' effectiveness relies heavily on the data they analyze.

Vectra AI Pro addresses this by merging network, identity, cloud signals into a unified view of an attacker's actions, which is then delivered to the agents for reasoning. The product also incorporates telemetry from software-as-a-service applications, endpoint detection and response tools, and secure access service edge infrastructure. Vectra AI dubs this collection of data as "trusted signal intelligence."

The intelligence covers all stages of the attack lifecycle, from reconnaissance and credential abuse to lateral movement. Vectra AI tracks data exiting the network in the same manner. Attribution follows, linking activities to users, workloads, or AI agents, and the company claims this extends to unmanaged systems that may not have been inventoried by a security team.

A third layer prioritizes entities and attack paths based on their significance. Vectra AI Pro includes a feature that enables an agent to demonstrate whether a path has been closed or an active attack has been stopped. Customers can adopt Vectra AI Pro in three different ways. They can integrate Vectra AI Pro with their custom agentic security operations center agents and copilots using REST application programming interfaces, a Model Context Protocol server, and prepackaged AI skills featuring indicator enrichment, PCAP retrieval, forensic investigation, hunting, and reporting.

Alternatively, Vectra AI provides its own agents that triage and prioritize detections, generating natural language summaries explaining the rationale behind flagging an entity. The third option is hiring Vectra's team to incorporate the signals into the customer's existing tools.

Vectra AI frames the launch in the context of a post-Mythos world, a reference to the Anthropic PBC model that has fueled discussions surrounding AI-enabled attacks. Hitesh Sheth, Vectra AI's president and chief executive, emphasizes that Vectra's key advantage lies in its ability to apply AI to understand attacker behavior, rather than just utilizing AI.

Security operations manager at Gamma Communications plc, AJ Wiggan, highlights how correlated signals help his team streamline investigation and response processes. He attributes this efficiency to the context Vectra AI Pro supplies about each incident. The product is now available for purchase, and Vectra AI is showcasing it at Black Hat USA in Las Vegas.

The company has raised approximately $350 million since its inception in San Jose in 2011 and holds 39 patents. Vectra AI serves over 2,000 organizations and has acquired network observability startup Netography Inc. last October, followed by a platform-wide detection and response framework in December.