The AI Governance Practitioner Program: Specialty Courses

AI Governance Specialty Domains

The Foundation Track builds the cross-domain capability every AI governance practitioner needs. The specialty courses build on that foundation - going deep into domains of expertise. Each specialty approaches AI governance through a different lens, with its own practical toolkit and a path to formal recognition through a Practitioner Award.

The courses are in development and well advanced with the first being released in July 2026. We expect the remainder to be available before the end of 2026. Join the waitlist for those specialties that matter most to you and we'll notify you when they're ready.

AI Compliance

AI Risk

AI Engineering

AI Evaluation

AI Operations

AI Leadership

AI Compliance Specialty

From regulatory expectations to unified controls and governance mechanisms. 

Many compliance courses in AI governance today cover little more than framework knowledge to transcribe: copying regulatory text into checklists and calling the result "compliance." This course teaches translation: understanding what requirements actually demand and designing controls that produce compliance evidence as a byproduct of well-functioning governance mechanisms.
The course is anchored in three primary regulatory sources: the EU AI Act, ISO 42001, and NIST AI RMF, but teaches a methodology that extends to any regulation, standard or other source of requirements. Through a running case study of a fictional firm navigating all three simultaneously alongside GDPR and client requirements, you progressively build essential design artefacts: a crosswalk map showing how external expectations from multiple sources map onto a unified set of internal controls, and a mechanism portfolio demonstrating how controls are implemented through functioning closed-loop mechanisms.
The method used throughout is: Artefact  Expectation  Control  Mechanism. You learn a unified control framework across twelve governance domains, the discipline of parsing expectations from regulatory text, crosswalk construction, and mechanism design using a seven-component diagnostic. The method works regardless of which regulation, standard, or framework you face. You'll also learn to use a brand new toolset for regulatory mapping and mechanism design.
The Compliance Speciality course has 140 video topics, plus six guided exercises and 700 quiz questions. It will take between 60 and 70 hours of learning time to complete.  

Release Date:

July 2026

Who is this for?
Governance, legal, risk, compliance, and assurance professionals. Anyone responsible for translating regulatory requirements into functioning controls.
Join the AI Compliance Waitlist

    AI Risk Specialty

    Making AI risk management operational and responsive

    This specialty builds the capability to understand where risks surface across AI system components, and then builds the operational mechanisms that make risk management continuous and adaptive.
    The course builds analytical capability through progressive threat modeling of an AI system that evolves in complexity. You start with a basic model and trace risks across its components, then the system grows: it gains autonomy, integrates external data, connects to tools, scales to production, and becomes agentic. At each stage, new risks emerge and you learn to identify them, distinguish static risks from dynamic ones that surface only through operation and interaction, and select controls proportionate to the threat. 
    From that foundation, the course builds operational capability through three mechanisms that make risk management a continuous function. These include a combine mechanism for risk identification, assessment, and treatment within a cohesive workflow; keeping risk management alive through continuous monitoring and governance cadences; and ensuring that changes, whether driven by incidents, regulatory shifts, or system updates, feed back into the process before they become unrecognized risks.
    The Risk Speciality course has 117 video topics and 585 quiz questions. It will take between 55 and 60 hours of learning time to complete.  

    Release Date:

    Waitlist Driven

    Who is this for?
    Governance, risk, legal, audit, product and security professionals. Anyone who needs to understand how AI-specific risks differ from traditional risk.
    Join the AI Risk Waitlist

      AI Engineering Specialty

      Designing operational governance into agentic AI architecture. 

      Safety and security are most effective when they are properties of a system's design, not controls added after the architecture is set. This specialty course teaches the engineering discipline that makes governance structural, so that unsafe or insecure behavior is prevented by architecture rather than caught by review.
      This is not a course on how to build AI systems. It teaches the engineering mindset and principles for designing safety and security into complex AI systems. The course is structured around six design rules that apply to every design decision in a system with autonomous capabilities: separate the control from the thing it constrains, verify everything that crosses a boundary, never rely on a single control for a safety-critical property, design every component for how it fails, ensure every action is observable and attributable, and ensure every control has a feedback signal that drives adaptation.
      Four recurring scenarios run throughout: agentic systems, RAG systems, ML pipelines, and multi-agent workflows. Topics include identity and delegation architecture, trust boundaries, defense in depth, agent loop safety, tool design, human oversight engineering, adversarial defense, observability architecture, evaluation gates, supply chain security, and failure design. Each topic uses counter-examples: the wrong design is shown, the rule violation identified, and the corrected design demonstrated.
      The Engineering Specialty course has 114 video topics and 570 quiz questions. You should expect it to take between 55 and 60 hours of learning time to complete.

      Release Date:

      Waitlist Driven

      Who is this for?
      Engineers, architects, and the governance, risk, and audit professionals who need to evaluate whether safety and security are designed into AI systems.
      Join the AI Engineering Waitlist

        AI Evaluation Specialty

        The measurement discipline every governance function depends on.

        Every governance discipline depends on evaluation, and if your evaluation is weak, everything built on top of it is unreliable. 
        The course is structured around eight questions that form a practitioner's evaluation reasoning chain: what am I evaluating, what should I be looking for, how do I design tests that reveal what I need to know, how do I measure what I find, how do I stress-test it, how do I know whether to trust my results, how do I read someone else's results, and how do I keep knowing. The arc moves from doing evaluation, to validating it, to sustaining it over time.
        Three recurring scenarios ground the concepts in practice: a RAG-based knowledge assistant, a customer-facing agent with tool access, and a hiring classifier. The course covers scoping, test design for non-deterministic systems, metrics and their limitations, adversarial evaluation including red teaming and OWASP and MITRE frameworks, epistemic rigor, critical interpretation of benchmarks and vendor claims, and continuous evaluation design.
        The Evaluation Specialty course has 102 video topics and 510 quiz questions. You should expect to take between 45 and 55 hours of learning time to complete.

        Release Date:

        Waitlist Driven

        Who is this for?
        Anyone who designs, commissions, interprets, or makes decisions based on AI system evaluations. Technical and non-technical practitioners alike.
        Join the AI Evaluation Waitlist

          AI Operations Specialty

          Building the platform and practices to govern AI systems in production.

          Governance that has been designed and assessed still needs to function in production. This specialty course teaches how to build and run the operational machinery that keeps AI systems governed, not as a one-time setup but as a continuous, adaptive, inspectable system.
          The course builds an operational governance platform using an open-source stack: governance, risk and compliance records (VerifyWise), machine learning operations (MLflow), evaluation evidence (DeepEval), data quality (Great Expectations), production monitoring (Evidently), policy enforcement (OPA), runtime guardrails (NeMo Guardrails), workflow orchestration (n8n), and conversational governance interface (MCP). You learn what each component does and how they connect as a governance platform.
          From that foundation, the course designs six operational governance mechanisms: deployment governance, production monitoring and response, incident detection and response, data governance, model lifecycle governance, and continuous compliance evidence. Each mechanism is worked through as a complete design built around the governance control loop: Sense, Decide, Constrain, Actuate, Evidence.
          The Operations Specialty course has 102 video topics, six embedded worked examples, and 510 quiz questions. You should expect it will take between 50 and 55 hours of learning time to complete.

          Release Date:

          Waitlist Driven

          Who is this for?
          Operations and platform engineers, and the governance, compliance, and audit professionals who need to understand how controls function in production.
          Join the AI Operations Waitlist

            AI Leadership Specialty

            Leading AI governance programs from business case to sustained culture

            AI governance presents challenges that cannot be solved with technical expertise alone. Building the business case, cultivating a governance culture, navigating organizational resistance, sustaining commitment through leadership transitions. These are adaptive challenges that require adaptive leadership. 
            The course builds five leadership responsibilities that determine whether governance succeeds or fails. You learn how to translate organizational values into principles and then into measurable commitments that hold people accountable. You learn how to design governance into how the business actually runs, rather than layering it on top. You learn how to cultivate a governance culture intentionally, how to steer a portfolio of governance mechanisms as they mature, and how to create the conditions for others to exercise leadership across the organization. 
            Over 50 case studies of successes and failure drawn from Waymo, Cruise, the WHO Surgical Safety Checklist, Virginia Mason, Johnson & Johnson, OpenAI, and Anthropic show what these responsibilities look like in practice, and what happens when they're absent.
            The Leadership Specialty course has 134 video topics, embedded diagnostic exercises, and 670 quiz questions. You should expect to spend between 60 and 70 hours of learning time to complete.

            Release Date:

            Waitlist Driven

            Who is this for?
            Senior executives, program leads, consultants, legal counsel, auditors and anyone responsible for creating the conditions in which AI governance succeeds.
            Join the AI Leadership Waitlist
              Ways to Learn

              Self-paced or a guided practitioner cohort.

              Self-paced Online

              Work through the material at your own speed. Every topic includes videos, quizzes, and exercises to test your understanding as you go. On completion, you earn a Certificate of Completion.

              Guided Practitioner Cohort with Live Sessions

              A small group of up to 15 practitioners, guided over eight weeks by James Kavanagh. Complete the graded assessment at the end to earn a Specialty Practitioner Award - which counts toward the Master Practitioner pathway.

              Frequently asked questions

              Why are the release dates "waitlist driven"?

              We're building the specialty courses in the order that demand tells us to. Rather than deciding which course to prioritize based on assumption, we're letting practitioners vote with their interest. The specialties with the most waitlist registrations will be developed and released first. If you want to influence what gets built next, joining a waitlist is the most direct way to do that.

              Can I join more than one waitlist?

              Yes. If more than one specialty is relevant to your work, join as many as you like. You'll be notified separately when each one is ready.

              When will the first specialty course be available?

              AI Compliance will be available in July 2026 but enrollments will open in mid June. For the remaining five, release timing depends on waitlist demand. We'll notify everyone on a waitlist as soon as a course is confirmed for development and again when it's ready to enroll.

              Do I need to complete the Foundation Track before taking a specialty course?

              The specialty courses are designed to be taken after the Foundation Track. They build on the shared language, adaptive governance methodology, and cross-domain foundations that the Foundation Track establishes. Without that foundation, some of the specialty content will be harder to apply in practice.

              I already have significant experience in my domain. Do I still need the Foundation Track?

              The free AI Governance Practitioner Capability Assessment will tell you where your capability sits and whether the Foundation Track is the right starting point for you, or whether you're ready to move directly to a specialty.  But the Foundation Track teaches the methods of governance mechanism design that you will likely still want to go through in depth.

              What's the difference between a Certificate of Completion and a Specialty Practitioner Award?

              A Certificate of Completion confirms you worked through the course material. A Specialty Practitioner Award confirms your capability was tested through a graded assessment. Participating in a Practitioner Cohort includes the Assessment. Only the Practitioner Award counts toward the Master Practitioner pathway. 

              Do I need to take specialties in a particular order?

              No. Each specialty is self-contained. You can take them in any order, depending on where your work takes you.