Guild Innovation

Proprietary frameworks. IBM underneath.

Visitors should not have to infer that Guild is a technology company. AEGIS, ControlPlane, token economics, and quantum-safe cryptography are first-class offerings — implemented with watsonx, OpenPages, Instana, and DataPower.

Project Guild ControlPlane

AI Governance Control Plane

Full AI lifecycle governance, from intake through retirement.

A lifecycle on a slide is not an operating model. ControlPlane is the connected view a CIO can use: inventory, ownership, lineage, control, cost, evidence, and value — current, not another dashboard.

The problem

Enterprises now have approved, experimental, embedded, third-party, and shadow AI. Identity, risk, policy, controls, evidence, telemetry, and enforcement have to travel with the system, or governance is theatre.

The model

Intake and classification of every AI system that can act
Named owners, recertification, exceptions, and retirement
Lineage across data, models, prompts, tools, APIs, and downstream systems
Runtime policy, human approval, and a tested kill switch
Evidence that can be shown to audit, risk, and the board

What the client gets

Governed AI inventory
Lifecycle control model
Runtime observability
Evidence bundles
Re-risk triggers
Kill-switch / intervention paths
Cost attribution

Outcome. A CIO can see every AI system that can act — who owns it, what it can do, what it costs, and whether it should still be running.

Cost that can be forecast

AI FinOps / Token Economics

Usage, cost, ROI, and predictive budgeting across tools and agents.

Model spend, token volume, GPU, and human oversight are now a line item. Guild treats AI cost as an operating discipline — mapped to Turbonomic and Apptio where the rest of the estate already lives.

The problem

Teams can stand up an agent in a week and then fail to answer what it costs, who pays, and whether the outcome is worth the run-rate.

The model

Inventory of models, tools, and agents with a named cost owner
Token, inference, and infrastructure unit economics
Budgets and alerts before the invoice is a surprise
Value evidence next to spend — not a separate slide

What the client gets

Named cost owners on models, tools, and agents
Token, inference, and infrastructure unit economics
Budgets and alerts before the invoice is a surprise
Value evidence next to spend

Outcome. AI spend is forecastable, attributable, and defensible — on Turbonomic and Apptio where the rest of the estate already lives.

Guild AEGIS

Agentic AI Framework

Identity, authorization, observability, and lifecycle for enterprise agents.

Guild AEGIS — AI Enterprise Governance & Intelligence System — is how we implement agents that can act, with an architecture that can stop them. Building the agent is not the hard part.

The problem

An agent that can call tools, change records, or spend money without identity, policy, and evidence is not an enterprise system. It is an ungoverned intern with production credentials.

The model

Agent runtime: perceive, reason, plan, remember, act, validate, respond
Control plane: identity, registry, entitlements, policy, human approval, kill switch
Observability: who acted, on what data, with which tool, under which policy
Lifecycle: intake, promotion, recertification, and retirement

What the client gets

Agent identity and registry
Entitlements and runtime policy
Human approval on consequential actions
Tested kill switch
Observability: who acted, on what, with which tool

Outcome. Agents that can act in production with an architecture that can stop them.

Crypto inventory → PQC migration

Quantum Security Framework

Inventory, quantum-risk assessment, prioritization, and algorithm-agile migration.

Harvest-now, decrypt-later is a present risk. NIST has published ML-KEM, ML-DSA, and SLH-DSA. Guild inventories TLS, signing, gateways, and PKI, then migrates in phases on the IBM estate — DataPower, MQ, Guardium, OpenPages.

The problem

Certificates, VPN, MQ, API gateways, code signing, and long-lived archives are already on a clock. Most estates do not have a cryptographic bill of materials, let alone a sequenced migration with owners and exceptions.

The model

Cryptographic bill of materials across applications, gateways, partners, and HSMs
Risk ranking by data longevity, algorithm, and blast radius
Hybrid TLS and certificates first, then signing, then archives
Crypto-agility as the operating model — not a one-time algorithm swap
Optional IBM Quantum pilot only where the problem actually needs it

What the client gets

Cryptographic bill of materials
Risk ranking by data longevity and blast radius
Phased hybrid / PQC migration plan
Crypto-agility as an operating model

Outcome. The estate is inventoried and sequenced before Q-day — on DataPower, PKI, and Guardium, not as a qubit demo.

Review an accelerator against your estate

ControlPlane, AEGIS, token economics, or quantum-safe — start with the operating model, not a product demo.

Review AI governance architecture