About SyzygySys — Credo, Ethos, Mission

Credo

What We Believe

AI agents are now first-class contributors to World 3 — Karl Popper’s world of objective knowledge that exists independently of any individual mind. They produce theories, decisions, and artefacts at machine speed. That is not a metaphor. It is a description of what is happening in production infrastructure today.

No one currently governs the identity, access, or agency of this class of entity in any systematic, portable, or auditable way. The entire security and governance stack was built assuming a human is the actor at the end of every chain. Agents are not human. The assumptions do not hold.

Trustworthy systems must be buildable from untrustworthy components. This is not a philosophical aspiration — it is the foundational engineering requirement of the agentic era. You cannot trust the model. You cannot trust the tool. You cannot trust the agent’s self-reporting. The governance layer must not depend on the trustworthiness of any individual component. (NCSC Problem Book, Cross-Cutting Problem #1.)
Security belongs inside the framework, wrapped around inference, not bolted onto the outside. Runtime enforcement, identity at the point of action, policy that travels with the agent — these cannot be retrofitted onto an architecture that was not built for them.
Whoever controls the identity, access, and agency of AI agents shapes how the agentic era unfolds. This should not be one vendor. It should not be one cloud. It should be the enterprise, the operator, the human principal — with cryptographic certainty, portably, across any environment.
Inference that is ungoverned is a liability. Inference that is governed is a capability. The same agent that is a risk under no governance is an asset under the right governance. The difference is the control plane.

Ethos

How We Work

Zero Debt

Zero technical debt. Zero security debt. Not as a target — as a standing condition. We do not accumulate what we cannot defend.

We Eat What We Make

ACE governs its own development. The platform manages its own agents. Every claim we make about governance is made by a system that has been governed by that claim in production.

Measurable or It Doesn’t Count

77% test coverage on core platform, approaching 100% with documented exceptions. Error logs at or near zero per day, annotated when they occur. Policy enforcement rate, HITL trigger rate, denial rate — all quantifiable from the audit trail.

Mindful of the Caveats

We know what ACE does not yet do. We document it. The things outside the current specification’s scope are named, not hidden. Honest scope is a security property.

Open to Scrutiny

We welcome external audit. We publish our research. We open-source our developer tools. We invite challenge. A platform built for transparency should not be opaque about itself.

Recursive Integrity

The system builds itself. The agents that built ACE are governed by ACE. The infrastructure that runs ACE is managed by ACE. This is not a demo. It is how we work.

Mission

What We Are Here to Do

To be the vendor-agnostic agent control plane.

“A third plane is now emerging. As enterprises deploy multiple agents from multiple vendors across many domains, a new question becomes unavoidable: how do we apply a consistent envelope of visibility, governance and management across a heterogeneous agent estate. This is the role of the agent control plane. Distinct from a development environment or an orchestration fabric, this plane supervises the entire agent heterogeneous landscape in a vendor-agnostic way to keep autonomous behavior aligned with business intent, policy and risk tolerance.”

Leslie Joseph, Forrester — Announcing Forrester’s Research: Agent Control Planes, March 2026

That is ACE. We built it before Forrester named it.

PORTAL · VIGIL · RAT · HITL · DACP

Govern the identity, access, and agency of AI agents — across any orchestrator, any cloud, any identity provider — with portable cryptographic identity, pre-execution policy enforcement, tamper-evident audit, human gates on high-risk actions, and a governed agent mesh that prevents lateral movement.

NCSC Problem Book

Close the gap the NCSC named as unsolved: building trustworthy systems from untrustworthy components (Problem #1) and creating meaningful measures of cyber security (Problem #3) — for the new class of autonomous actor operating inside enterprise and national infrastructure.

Vendor-Agnostic

Ensure the independent alternative exists. Microsoft is building a proprietary portable agent identity registry. Every major cloud vendor is building proprietary governance layers. Enterprises that run multi-cloud, multi-orchestrator, or non-Microsoft agent stacks need an independent standard. ACE is that standard.

Human-in-the-Loop

Keep humans in the loop — not as a constraint on capability, but as a design principle. Trust is earned through telemetry. Trust is bounded by policy. Trust is conditional on auditability. On high-risk actions, a human approves before the agent acts. That will not change.

Team

The People Behind ACE

Kevin Broderick

Founder & Chief Architect

Platform engineering and security architecture background. Designed and built all 16 pillars of the ACE platform — from CROWN routing and VIGIL boundary inspection to LEDGER provenance chains and DACP lifecycle governance. Responsible for architecture, engineering, security posture, and regulatory alignment across the full stack. 206K lines of production code across 44 projects. 77% test coverage on core platform. Zero technical debt. Zero security debt.

View CV

Foundation

SyzygySys Ltd

Company

SyzygySys Ltd
CRN 16967021
Incorporated in England and Wales

HMRC SEIS

Advance Assurance granted
2026-03-03
Ref WMBC/I&R/28386 29344/VCRT

Compliance

GDPR privacy by design
EU AI Act self-assessed
Fully open to external audit

Production

206K LOC · 44 projects
30+ running microservices
0 tech debt · 0 security debt

We built the agent control planebefore Forrester named it.