Implement auditable AI governance for consequential systems

We do not sell generic AI-governance slide decks. We help organizations implement the technical and operating controls needed to demonstrate that consequential AI actions were authorized, policy-bound, correctly executed, and independently verifiable.

Core offers and deliverables

1. AGEI Readiness Assessment

Offer: Determine what must change for the client to move from policy documents and fragmented operational records to an auditable governance-evidence capability.

Deliverables:

• AI/agent inventory and prioritized use-case map
• Current-state governance and technical-control assessment
• Authority, access, and sensitive-data flow map
• Evidence-gap analysis: which controls they can substantiate with evidence, versus claims that rely on narrative, screenshots, or mutable records.
• Agent/tool risk classification and control tiering
• Shadow AI exposure and discovery recommendations
• Target AGEI architecture and minimum auditable path
• 90-day implementation roadmap
• Executive readout and technical working session

Scope boundary: This is an assessment and architecture engagement, not a full technology deployment. Best for organizations with active AI or agent initiatives that need a technical baseline before selecting controls, platforms, or implementation priorities.

Commercial approach: AGEI Readiness Assessments are scoped after an initial Readiness & Scoping Conversation. Typical engagements begin at $15,000 and vary based on the number of AI or agent workflows, business units, regulated data, systems, integration requirements, and assurance objectives.

2. Governed Agent Design Sprint

Offer: Design the authority and control model for one prioritized agentic workflow before it gains consequential operational access.

Deliverables:

• Agent role, purpose, and decision-boundary specification
• Tool inventory and risk classification
• Delegated-authority model
• Tool permissions, resource scopes, and action limits
• Attribute-Based Access Control (ABAC) / Role-Based Access Control (RBAC) and just-in-time elevation design
• Approval and escalation matrix
• Segregation-of-duties design for sensitive actions
• Policy-as-code requirements and rule catalog
• Revocation, session expiry, and kill-switch design
• Threat model covering indirect prompt injection, tool misuse, privilege escalation, replay, and bypass paths
• Architecture diagram and implementation backlog

Example: procurement workflow. A procurement agent may draft a purchase request autonomously, submit an approved request below a set limit, require a named approver above that threshold, and never gain direct access to payment execution.

Price: $25,000–$40,000, depending on whether the work covers one workflow, a reusable authority-control pattern, or multiple agent classes.

3. Proof-Carrying Execution Pilot

Bounded agent actions verified before execution.

Offer: Build one working, bounded implementation that proves the AGEI pattern in the client’s own stack.

Deliverables:

• One governed agent workflow operating in a controlled environment, which may include a narrowly scoped production deployment where appropriate.
• Agent identity, session, and delegation model
• Policy gate with approve, deny, escalate, and inspect outcomes
• Action-bound, short-lived, signed proof-bundle design
• Tool gateway or adapter that verifies proof before execution
• Parameter validation, target-resource binding, and replay prevention
• Request, decision, approval, and execution-outcome receipts
• Evidence graph, verification workflow, and sample audit pack
• Test suite for denied actions, expired proofs, altered parameters, replay attempts, and revoked authority
• Runbook, threat model, architecture documentation, and handoff workshop

Price: $50,000–$85,000 for one bounded workflow and one or two integration targets.

4. Enterprise Evidence Integration

Offer: Turn the pilot pattern into a reusable organizational capability.

Deliverables:

• AGEI reference-schema mapping to existing systems
• IAM, SSO, service-identity, and delegated-credential integration
• SIEM and security-event forwarding
• GRC, ticketing, and approval-workflow integration
• Model registry, evaluation, and lifecycle-evidence integration
• Policy and gate-management workflow
• Key-management, signing, rotation, revocation, and verification design
• Audit-pack templates and evidence-retention procedures
• Multi-workflow onboarding playbook
• Control-owner operating model and RACI
• Conformance-test suite and implementation scorecard
• Executive roadmap for scale-out

Commercial structure: Typically phased, beginning with the highest-risk workflows and control integrations.

Starting range: $90,000–$175,000+, depending on systems, workflows, and assurance requirements.

5. Fractional AI Governance & Assurance

Offer: Ongoing technical governance leadership for organizations that have AI initiatives but do not need or cannot yet hire a full-time governance architect. This engagement provides technical governance leadership and implementation oversight; it does not replace the client’s legal counsel, compliance accountability, security ownership, or executive decision authority.

Monthly deliverables:

• Monthly governance-control review
• Review of new AI use cases, agents, tools, and integrations
• Policy and delegated-authority design support
• Evidence-pack and audit-readiness review
• Shadow AI trend and remediation review
• Risk register and control-status updates
• Executive or board-ready briefing
• Implementation office hours for engineering, security, privacy, and legal teams
• Quarterly maturity assessment and roadmap refresh

Commercial terms: Limited number of stakeholders/workstreams, with major engineering implementation scoped separately.

Price: Starting at $8,000 per month (three-month minimum). For high-stakes or multi-team environments, target $12,000–$15,000/month.

Specialist modules

Specialist modules are designed as add-ons to the core engagements above.