Blueprint: Build the Best in Cyber Defense

Building Trust Into Agentic SOC Tools with Oren Saban

22 snips
Jul 2, 2026
Oren Saban, Chief Product Officer at Mate Security and ex-Microsoft Security Copilot team member, builds AI for SOCs. He explains why agentic SOCs use swarms of specialized agents. He discusses data quality, confidence calibration, which analyst skills matter, and the risk of prompt injection attacks. Short, sharp conversation on trust, testing, and practical tradeoffs when AI leads investigations.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Agentic SOCs Deliver Real End to End Workflows

  • Agentic SOC platforms now perform end-to-end triage, investigation, threat hunting, reporting, and automation beyond previous AI/ SOAR attempts.
  • Oren Saban contrasts older SOAR limits with agents leveraging unstructured knowledge to automate exception-heavy enterprise cases reliably.
ADVICE

Curate Context Before You Feed Agents

  • Don’t blindly ingest every doc; build data sync, curation, and confidence levels before trusting Confluence, Slack, or Teams content.
  • Oren Saban recommends structuring and scoring contextual data so agents don't leverage stale or garbage information.
INSIGHT

Specialized Agent Swarms Beat One Generalist Model

  • Splitting investigations into specialized agents improves determinism, security, and quality versus one generalist model.
  • Oren explains scoped agents reduce context load, enable least-privilege for sensitive files, and raise task accuracy.
Get the Snipd Podcast app to discover more snips from this episode
Get the app