Sekoia ships Elevate autonomous SOC agent after 425K tests
Sekoia announced the general availability of Elevate, an autonomous security operations center (SOC) agent, on September 30, 2026, from Paris, France. The product is positioned as the "agentic AI layer" of Sekoia's autonomous SOC platform and represents the company's first shipping agent product for enterprise security teams.
How Elevate works
Elevate operates as an end-to-end autonomous investigator. When a security alert fires, the agent ingests customer security data, conducts investigation workflows, and produces an audit-ready verdict complete with a traceable record of queries and evidence. Sekoia reported an average verdict time of 99 seconds—critical for SOC teams operating under time pressure during active threat response.
The verdicts are human-auditable by design. Each investigation includes documented reasoning and evidence chains, allowing analysts to understand why the agent reached its conclusion and to override or refine decisions if needed.
Scale and validation
Before GA, Sekoia ran an early-access program in production customer environments. The program involved more than 2,000 customers, hundreds of SOC analysts, and more than 425,000 autonomous investigations. This deployment at scale—rather than closed beta or internal testing—signals confidence in the agent's reliability in live SOC workflows.
Reuters reported that Sekoia rolled out Elevate to its clients after the early-access program and described it as a cybersecurity AI-agent tool, confirming the shift from pilot to general availability.
Market context
The SOC automation space has attracted significant vendor attention, but most offerings remain alert-routing or ticketing automation. Elevate attempts to push deeper—automating the investigation verdict itself, not just alert triage. The 99-second average reflects a design goal to keep human analysts in the loop for complex decisions while accelerating routine threat assessment.
The 425,000-investigation dataset from the early-access program provides Sekoia with rare production data on agent performance across real security environments, threat types, and customer infrastructure—a competitive moat difficult for competitors to replicate quickly.