OMDIA TECHNICAL SUMMARY | 2026 edition

Autonomy without losing accountability

The 2026 Omdia Technical Summary on governed agentic SecOps.

Security teams are absorbing a massive surge in data, and adding headcount isn't a sustainable answer. The real question: can you run autonomous, machine-speed investigation and keep it governed, explainable, and measurably effective at scale?

In a new 2026 Technical Summary, Omdia reviewed Foresite's operations data on Google SecOps and found that practitioner-governed AI delivers measurable outcomes, not just hype.

UNGATED FINDINGS

The proof is in the data

Measured by Omdia across Foresite's live operations on Google SecOps.

~60%

drop in mean time to investigate

Machine-speed analysis cuts response times.

278%

telemetry surge, investigation load flat

Capacity is decoupled from data volume, the system scales analyst effectiveness against growth.

25→86%

benign-event auto-identification

As ticket volume nearly tripled, analysts stopped chasing harmless events. Accurate, not just fast.

HOW IT WORKS

The glass-box Triage Agent

Autonomous investigation runs at machine speed, but the actions you can't take back are validated by a named practitioner before they execute. The Catalyst Triage Agent presents a confidence-scored verdict with its evidence and reasoning exposed. Every high-impact action keeps a human in the loop, an agent you can govern, not one you're forced to trust.

01 Machine-speed investigation across Google SecOps telemetry

02 A confidence-scored verdict, with evidence and reasoning exposed

03 Named-practitioner validation before any irreversible action

Autonomy without losing accountability

The proof is in the data

The glass-box Triage Agent

Want the deep dive into the architecture?