European cybersecurity company Sekoia has made the AI layer of its autonomous security operations centre (SOC) platform generally available, claiming it can investigate security alerts and deliver a verdict in an average of 99 seconds.

The new platform, called Sekoia Elevate, uses AI agents to investigate alerts by analysing security data, threat intelligence, logs, detections and response information before presenting security teams with what the company describes as an audit-ready conclusion.

Sekoia said the technology is designed to automate the investigative work that is typically conducted by analysts and will enable them to spend more time on reviewing findings and making decisions.

The announcement follows an early access programme in which Elevate was deployed across more than 2,000 customer environments. During the trial, the platform completed more than 425,000 autonomous investigations.

According to Sekoia, managed security service providers using the technology reported reducing overall investigation times by up to five times while improving the consistency of threat analysis.

The company said transparency was a key part of the platform. Each investigation records the queries run by the AI agents and the evidence used to reach a conclusion, allowing analysts to review, challenge or override findings. Any changes are logged to support auditing and incident reviews.

Sekoia added that each completed investigation helps improve future investigations by adding validated information to the platform's knowledge base.

The platform has been built to work with existing security tools rather than requiring organisations to adopt a single technology stack. Sekoia said Elevate uses a model-agnostic approach, selecting different AI models depending on the task, while drawing on its own threat intelligence, security data infrastructure and integrations.

Georges Bossert, chief technology and product officer at Sekoia, said: “Elevate shifts the analyst's role from data miner to decision-maker. Our agents do the exhaustive groundwork from raw data and explain every step, so security teams can focus on the decisions that protect the business, without giving up control.”

In recent months, AI has made headlines for its involvements on the other side of cyber security. Frontier models from leading developers including OpenAI and Anthropic have breached numerous governments during testing, leading key figures in the industry to call for slowdowns in model development.

---