Guild ControlPlane · Control Tower

GUEST

Prototype environment — all costs, events, users, and outcomes are simulated.

Guild SystemsBoxAI Governance Control Plane · Generic Case StudyIBMLive · reference framework

AI Governance Control Plane

From documented policy to provable control

A generic enterprise case study, a live reference framework, and a pitch for why ControlPlane was built this way. Documentation alone cannot prove control — only an enforced, observed, and evidenced pipeline can. This module walks the five governance domains, the end-to-end golden trace, and the program outputs that turn policy into regulatory proof.

Control Status Summary

11 of 14 controls enforced in-band

79%enforcement coverage
Active
11 / 14
Completed
0 / 14
Pending
3 / 14
Total Controls
14
Sections

01 · Entry

Governance Maturity Model

Where the active client sits today, derived from the live simulation. The goal is Level 5 — continuous governance assurance.

...

02 · Architecture Vision

Business Policy → Regulatory Proof

A 7-layer pipeline from written policy to packaged regulatory proof, and the four enterprise outcomes it enables. Click any node for detail.

...
...
...
...
...

03 · Control Plane Domains

Five Domains of AI Governance

Identity & Lifecycle, Governance Gate, Telemetry & Evidence, Tools & MCP Access, and Audit & eDiscovery — each with its core principle and mechanism catalog (CP-01 through CP-14). Click a card to expand description, decision focus, example, and best practice.

Domain 1 · Identity & Lifecycle

Every agent has a verifiable identity, a declared owner, and a managed lifecycle from registration to retirement.

04 · Golden Trace Flow

An End-to-End Governed AI Action

Run the golden trace to watch a single action flow through identity, policy, tools, model, output, and evidence — each node proving what it must.

Golden Trace · End-to-End Governed Action

User → Identity → Agent → Policy → Tool → Model → Output → Outcome → Evidence

IDLE
1

User

2

Identity & Authority

3

Approved Agent

4

Policy & AuthZ

5

Tool / Data

6

Model

7

Output Protection

8

Outcome

9

Evidence Repository

Golden Trace Log

IDLE

Press Run Golden Trace to stream the governed action.

evidence sealed · SHA-256 …e9b204

05 · Guided Workshop

Workshop Session Flow

Five sessions walked in sequence — the original workshop structure. Check off expected outcomes, review key decisions, and close each session.

Workshop Mode · Guided Session Flow

5 Sessions · in sequence

0/5 closed

Objective

Establish that every agent has a verifiable identity, a declared owner, and a managed lifecycle.

Validation question

Can you list every production agent, its owner, and its current lifecycle state right now?

Expected outcomes

Key architecture decisions

  • Adopt a single agent catalog as source of truth
  • Bind identity to credentials that expire
  • Make retirement a telemetry-emitting event

06 · Development Process

Governance Architecture Development Process

Seven steps to build the control plane, each producing a named output artifact — from the material AI action taxonomy to golden trace validation.

Governance Architecture Development Process · 7 Steps

Technical

Enumerate the actions the business treats as material (advisory, automated, customer-facing, PII-touching). Each gets a risk class and an owner.

Business

Focuses governance where it matters - not every API call is material, but the ones that touch customers or money are.

07 · Success Criteria

Program Outputs & Success Criteria

What a complete governance program produces: eight deliverables and a six-point success criteria strip — the proof the program is real.

...
...

08 · Traceability Model

From Requirement to Deliverable

The 8-layer traceability model: how an abstract governance requirement is realized, layer by layer, into a client deliverable and validation evidence.

Traceability Model · 8 Layers

How abstract governance requirements become implementable architecture and validation evidence. Each layer is the predecessor of the next — a requirement at the top is realized as a deliverable at the bottom.

#LayerTelemetry & evidence focus
1Enterprise RequirementWhy traceability is required
2Platform RequirementCorrelation, identity, governance capabilities
3Architecture DesignTrace propagation and event flow
4Security ControlPolicy enforcement and authorization
5Platform CapabilityTelemetry, evidence, and audit services
6Implementation ComponentGateway, runtime, policy engine, identity, repository
7Validation EvidenceGolden Trace, audit artifact, proof of enforcement
8Client DeliverableStandards, architecture, governance evidence model
Enterprise RequirementArchitecture DesignImplementation ComponentClient Deliverable

09 · Operating Model

Who Owns What

Governance never executes anything — the platform does. This matrix maps each responsibility area to the functional team that owns it, so every demand has an owner and an evidence path.

...

10 · Live Risk Playback

Risk Score Replay

Play back the simulation event stream and watch how the composite risk score moves across AI cost-attribution events — anomaly, policy gate, PII scan, prompt injection, and approval each contributing. Inject an anomaly or governance incident to see the score spike in real time.

...
Guild Systems

Guild ControlPlane

ControlPlane — AI Enterprise Governance & Intelligence System

Choose a Client Demo

Each client is a self-contained simulation context with independently seeded data, spend patterns, and governance rates.