Sekoia launches Elevate agentic SOC layer with 99-second verdict claim
Sekoia has announced the general availability of Elevate, the agentic AI layer of its autonomous security operations centre (SOC) platform. The Paris-based cybersecurity vendor says Elevate's AI agents investigate security alerts end to end, correlating raw logs, threat intelligence, detections and runbook context to produce a traceable, audit-ready verdict in an average of 99 seconds.
The product reaches general availability after an early access programme run across real production environments. Sekoia reports that more than 2,000 customers and hundreds of SOC analysts used Elevate during that phase, generating over 425,000 autonomous investigations. Managed security service providers (MSSPs) in Europe and the United States reported up to a fivefold reduction in overall investigation time, though Sekoia did not name specific MSSP customers or cite independently verified benchmark data.
How Elevate works
Each investigation produces a full audit trail: the platform logs every query its agents run and the evidence behind each finding, allowing analysts to review, correct or override any conclusion. Corrections are themselves logged and fed back into the system to improve subsequent investigations. Sekoia describes the architecture as model-agnostic, selecting the most appropriate AI model for each task rather than relying on a single foundation model, and says it is designed to integrate with a customer's existing security stack rather than replace it.
Chief executive Freddy Milesi said the company's view is that the underlying AI model is not the defining competitive advantage. "The model alone is not the moat," he said. "What makes an agent effective in the SOC is the system around it: threat intelligence, customer context, memory and the ability to act." Chief Technology and Product Officer Georges Bossert added that Elevate is designed to shift analysts from "data miner to decision-maker" by handling the exhaustive groundwork from raw data automatically.
Market context and competitive positioning
The autonomous SOC category is attracting substantial investment and competitive pressure. A number of well-funded vendors, including established SIEM and SOAR players as well as newer AI-native entrants, are pursuing similar alert-triage and autonomous-investigation capabilities. Microsoft, Palo Alto Networks and CrowdStrike have each integrated agentic or AI-assisted investigation features into their security platforms. Sekoia's positioning as a European-headquartered, model-agnostic, open-platform vendor is notable given growing enterprise and government sensitivity around data sovereignty, particularly as the EU's NIS2 Directive and the Cyber Resilience Act increase compliance obligations for security tooling across the region.
The EU AI Act also has direct relevance here. AI systems making consequential security decisions, even with human-in-the-loop override, may fall within the Act's high-risk classification depending on deployment context. Sekoia's emphasis on full traceability, analyst override and audit logging suggests a design posture aligned with those requirements, though the company has not publicly stated an EU AI Act compliance position.
What to watch
Near-term milestones for Sekoia will include the disclosure of named enterprise or MSSP customers, independently verified throughput and accuracy benchmarks, and any formal regulatory certifications such as SOC 2 Type II or ISO 27001 that would support procurement decisions in regulated industries. The company has not disclosed annual recurring revenue, total funding raised, or its current headcount, which limits external assessment of its commercial scale relative to competitors. Buyers evaluating autonomous SOC platforms will want to scrutinise false-positive rates and the governance controls governing agent actions, particularly for any automated response steps that go beyond verdict delivery into active threat containment.