Governing AI Agents in Production
Master the full governance lifecycle from strategy to compliance.
Outcomes
What you'll be able to do
Leave with production-ready systems, processes, and compliance frameworks you can deploy immediately.
Design enterprise governance systems
Build the reference architecture, inventory, and risk-scoring framework your organization needs to govern agents at scale.
Implement CI/CD and control gates
Deploy the paved road with automatic discovery, evaluation gates by risk lane, and release-scoped approvals that work in production.
Build observability and audit trails
Set up OpenTelemetry, collectors, golden metrics, and the evidence model that compliance and internal audit teams require.
Navigate EU AI Act and compliance
Understand the evidence model, risk levels, and documentation requirements for EU AI Act, GDPR, NIST AI RMF, and ISO/IEC 42001.
Syllabus
4 phases across 8 sessions
Why AI needs a new governance model
Chatbots answer, agents act; shadow AI and sprawl; the four jobs of governance; the business case and the three board numbers
Reference architecture on Azure
Hub-and-spoke landing zone; identity, network and model planes; five runtime archetypes (Foundry Agent Service, Container Apps, AKS, Functions, multi-agent); prompt-injection defenses
The lifecycle of an agent
Eleven governance stages vs five operational states; exit gates; three lines of defense; retirement
The agent inventory and risk scoring
Canonical inventory record and field authority; business and technical risk profiles; combined score, lanes and control mapping
CI/CD for agents
The paved road; automatic discovery; evaluation gates and thresholds by lane; HMAC-signed events; release-scoped approvals
AI governance as process
Intake, change and approval management; exceptions; multidimensional risk matrix; cascade re-approval across agents; incidents; RACI
Observability and operations
OpenTelemetry contract and SDK; collector as control point; eight golden metrics; alerting and drift; FinOps cost per agent
Audit, compliance and the end-to-end case
Evidence model and reconstruction; EU AI Act, GDPR, NIST AI RMF, ISO/IEC 42001, Colorado and UAE; ExampleCorp case study; control catalog
Audience
Who this program is for
AI Leaders & CTOs
Executives responsible for deploying agents safely at enterprise scale.
Compliance Officers
Teams ensuring agents meet regulatory requirements and internal policy.
Platform Engineers
Technical teams building the infrastructure and controls for production agents.
Resources
Learn more before you enroll
Watch the free lesson
30-minute Lightning Lesson with Jose Vicente Sáez and Daniel Andrés. Learn the core governance framework before the full program.
Course framework
The program follows the framework from Enterprise AI Governance on Azure by Daniel Andrés, José Vicente Sáez, and Javier Aguado.
Ready to govern agents at scale?
Join the waitlist and be the first to know when enrollment opens.