Course Objective & Scope
When an AI system in production makes a costly mistake or leaks data, governance fails if nobody knows whether accountability sits with the business unit, the data science team, IT or the vendor. This course teaches how to design an organisational role model that assigns unambiguous ownership across every stage of the AI lifecycle.
It covers adapting the Three Lines of Defence model to AI, positioning the CAIO or AI CoE, and defining role charters for system owners, technical leads and risk functions. In the workshop, participants build a tailored AI RACI matrix and escalation charter mapped to their actual organisational structure.
Who Should Attend?
- Executives, organisational design specialists and HR leaders
- AI Centre of Excellence (CoE), digital transformation and IT heads
- Enterprise risk, compliance, internal control and internal audit leaders
- Business unit heads taking operational ownership of AI systems
Key Takeaways
- Adapt the Three Lines of Defence model to enterprise AI governance
- Define clear charters for AI System Owner, AI Product Manager, Model Steward and Risk Owner roles
- Build a lifecycle RACI matrix from ideation through production monitoring and retirement
- Assign clear internal accountability even when using external vendors or low-code citizen AI
- Document decision rights, sign-off thresholds and escalation paths across committees
Curriculum
Closing the AI Accountability Gap with Three Lines of Defence
- Where ownership typically falls through the cracks in cross-functional AI initiatives
- 1st Line (Business & Engineering), 2nd Line (Risk, Compliance, Legal, InfoSec), 3rd Line (Audit)
- Why outsourcing a model or SaaS tool never outsources corporate accountability
Core AI Governance Roles and Role Charters
- Business / AI System Owner (accountable for operational outcomes and risk)
- AI Product Manager, Data Steward and Model / Prompt Owner
- AI Governance Lead, Data Protection Officer and AI Security Champion
Organisational Topologies: Centralised, Federated and Hub-and-Spoke
- Where to position the CAIO, AI CoE and AI Governance Committee
- Scaling oversight through embedded business-unit AI stewards
Designing the Lifecycle RACI Matrix
- Mapping decisions across intake, risk tiering, data clearance, testing, launch and monitoring
- Avoiding common RACI anti-patterns such as shared 'Accountable' assignments
Decision Rights, Veto Authority and Escalation Paths
- Defining sign-off authority by risk tier and emergency stop (kill-switch) rights
- Workshop: building the organisation's AI RACI matrix and role cards
Other Courses in This Category
-
Classify enterprise AI systems into risk tiers based on intended purpose, autonomy, affected stakeholders, data sensitivity and EU AI Act criteria, and map mandatory controls to each tier.
- Live Course
- 5 modules
-
Navigate the EU AI Act's extraterritorial impact, provider and deployer obligations, compliance milestones and penalties, and design a role-based corporate AI Literacy programme that satisfies Article 4.
- Live Course
- 5 modules
-
Manage KVKK and GDPR compliance in AI workflows, cross-border LLM API transfers, PII masking in prompts and RAG, copyright risks in training data, IP ownership of AI outputs, and protective contract clauses.
- Live Course
- 5 modules
-
Assess and govern third-party AI vendors, proprietary and open-weight foundation models, and external AI agents across data usage, security, model transparency, concentration risk and contractual SLAs.
- Live Course
- 5 modules
-
Design meaningful human oversight (EU AI Act Article 14) that avoids rubber-stamping and automation bias, using human-in-the-loop, human-on-the-loop, confidence routing and friction-aware review interfaces.
- Live Course
- 5 modules
-
Identify bias sources in machine learning and LLMs, select context-appropriate fairness metrics, and translate explainability methods (SHAP, LIME, citations) into clear disclosures for customers and regulators.
- Live Course
- 5 modules
Get in touch
