Evidence · autonomy · control · change

Security-agent claims, tested against the record.

Decision intelligence for enterprise security teams evaluating agentic systems, grounded in canonical evidence and confirmed by human review.

Confirmed intelligence

What changed

View all changes →
01

APPROVAL CHANGE

Project Perception documents human sign-off for every high-impact action, with defender-set objectives and guardrails.

2026-08-27

AUTHORIZATIONOPERATIONSRISK
02

ARCHITECTURE CHANGE

Project Perception documents a red/blue/green multi-agent architecture coordinated by orchestrated workflows and an orchestration harness.

2026-08-27

ARCHITECTUREEVALUATIONRISK
03

AUTONOMY FACET CHANGE

Project Perception assigns work execution to agents while reserving judgment to humans.

2026-08-27

AUTHORIZATIONEVALUATIONRISK
04

AVAILABILITY LIFECYCLE CHANGE

Project Perception is in preview, initially delivered through Microsoft Defender, with broader Microsoft Security expansion planned.

2026-08-27

DEPLOYMENTEVALUATIONPROCUREMENT

Coverage

Systems tracked

Compare systems →
MD

Microsoft

MDASH

Governed intelligence profile for MDASH

2 capability assertions · FRESH

PR

Microsoft

Project Perception

Governed intelligence profile for Project Perception

0 capability assertions · FRESH

OP

Open Security

Open Security Benchmark

Governed intelligence profile for Open Security Benchmark

0 capability assertions · FRESH

How ASI earns trust

Evidence in. Human judgment. Verified publication out.

01CollectGoverned primary sources
02GroundVerbatim evidence anchors
03ReviewMandatory human decision
04PublishValidated public projection

Evidence health

Open conflicts

No open conflicts in the current public projection.

Research cadence

Latest weekly analysis

Evidence pack · 2026-08-27