Artificial Intelligence for Leaders Training Course

Participants can frame AI opportunities as defined business problems with measurable value.

• Problem statements defined before selecting an AI solution approach
• Service journeys examined for delay, error, volume and decision friction
• Value hypotheses linked to revenue, cost, risk and experience measures
• Jobs-to-be-Done identifies customer and employee outcomes worth improving
• Use-case boundaries specify users, decisions, channels and exclusions
• Baseline measures established before forecasting AI-enabled performance improvements

Participants can rank AI initiatives using transparent value, feasibility and risk criteria.

• Weighted scoring models compare value, feasibility, risk and strategic fit
• Impact-effort matrices separate quick wins from long-term investments
• Data availability, quality and access tested during early prioritisation
• Dependencies mapped across platforms, processes, suppliers and scarce skills
• Portfolio roadmaps sequence experiments, pilots, scaling and retirement decisions
• Stage-gate criteria define evidence required for further investment approval

Participants can assess whether available data and performance evidence support an AI decision.

• Data quality dimensions cover accuracy, completeness, timeliness, consistency and validity
• Training, validation and test datasets separated to prevent misleading results
• Precision, recall and false-positive rates interpreted for operational decisions
• Drift detection identifies changing data, behaviour and performance assumptions
• Data lineage records source systems, transformations, access and retention
• Retrieval-augmented generation assessed for grounded responses using approved content

Participants can define safe, useful generative AI applications and their operating controls.

• Prompt patterns structure role, task, context, constraints and output format
• Hallucination risks assessed against factuality, source quality and decision impact
• Human-in-the-loop review designed for sensitive content and consequential actions
• Knowledge bases curated for retrieval-augmented generation and source traceability
• Content classification separates public, internal, confidential and restricted information
• Acceptable-use rules govern employee prompts, outputs and external sharing

Participants can identify responsible AI obligations and define proportionate safeguards.

• Fairness metrics examine disparate impact across relevant protected groups
• Explainability requirements matched to decision consequence and stakeholder needs
• Data Protection Impact Assessments identify privacy risks before processing begins
• Purpose limitation constrains data reuse beyond the stated business objective
• Accessibility and inclusion requirements considered within user and service design
• Documentation records intended use, known limitations and prohibited applications

Participants can establish an AI risk register and control approach for material threats.

• NIST AI RMF functions organise Govern, Map, Measure and Manage activities
• Risk registers capture cause, event, impact, owner, control and residual rating
• Prompt injection, data poisoning and model extraction threats assessed
• Red-team testing probes harmful outputs, bypasses and misuse scenarios
• Incident response procedures define detection, containment, notification and recovery
• Continuous monitoring tracks model performance, usage anomalies and control failures

Participants can assign decision rights, oversight roles and assurance requirements for AI systems.

• ISO/IEC 42001 management system principles structure AI accountability
• RACI matrices assign sponsorship, ownership, review and approval responsibilities
• Three Lines Model clarifies management, oversight and independent assurance roles
• Model cards document purpose, data, metrics, limitations and ethical considerations
• Approval authorities matched to risk tier, customer impact and regulatory exposure
• Audit trails preserve decisions, model versions, prompts and review evidence

Participants can evaluate AI suppliers and specify contractual protections for AI services.

• Vendor due diligence examines model provenance, data handling and subcontractors
• Requests for proposal define use cases, service levels and acceptance criteria
• Commercial models compare licence, consumption, implementation and exit costs
• Contract clauses address data ownership, confidentiality, audit rights and liability
• Service-level measures cover availability, response quality, support and incident reporting
• Proof-of-value criteria establish evidence before production-scale procurement commitments

Participants can prepare a controlled 90-day plan for an approved AI initiative.

• AI initiative charters define objectives, scope, sponsors, users and measures
• Change impact assessments identify affected roles, tasks, skills and controls
• Pilot designs set cohorts, hypotheses, guardrails and evaluation methods
• Training plans address user judgement, escalation and acceptable-use responsibilities
• Benefits realisation tracks agreed outcomes against baseline operational measures
• Ninety-day action plans sequence decisions, deliverables, owners and review dates