Artificial Intelligence

Overview

AI governance starts with what is already running

Organisations are often further into AI than their inventory says. Models arrive inside purchased software, teams adopt assistants without a procurement step, and a use that was a pilot last quarter is load-bearing this one. The first honest question is not which controls to apply but what is actually running, who owns it, and what it decides.

Closing that gap is mostly bookkeeping, and it is the part that gets skipped: an inventory that is maintained rather than compiled once, impact assessment that looks at the people affected and not only at risk to the organisation, and controls that survive a change of model or supplier. Most of it is governance rather than data science, which is why it usually lands on people who do not build models.

Professional tracks

Halderstone tracks follow a modular structure. They first establish a strong, role-specific foundation across disciplines, which is then applied to the chosen discipline.

Artificial Intelligence · ISO/IEC 42001ISO/IEC 42001 AuditorAuditing AI management beyond policies and principles◎Diploma in Management System Auditing◎Halderstone Certified ISO/IEC 42001 Auditor21 modules · capstone · examView the track →

Artificial Intelligence · ISO/IEC 42001AI ManagerTurn AI from an unmanaged risk to a managed capability◎Diploma in Management System Management◎Halderstone Certified AI Manager24 modules · capstone · examView the track →

Training modules

Each module can be taken on its own and trains the calls a practitioner has to make without being prompted. These are the modules that teach this discipline; the cross-discipline core that every track shares is in the catalogue.

HAM-AI-DF-02 · Artificial IntelligenceAI Limitations & Failure ModesBuild practical AI failure-mode literacy across predictive, generative and workflow systemsView details →

HAM-AI-S-02 · Artificial IntelligenceAI Risk ManagementAssess AI risks across use context, affected parties, GenAI and agency, then turn findings into treatment, monitoring and residual riskView details →

HAM-AI-S-01 · Artificial IntelligenceAI System Lifecycle & InventoryDefine AI system scope, set lifecycle boundaries, and maintain an AI system inventory aligned with ISO/IEC 42001View details →

HAM-AI-DF-01 · Artificial IntelligenceAI Systems & ArchitecturesPractical AI literacy for understanding AI types, deployment models, agentic patterns, current trends and where AI is headingView details →

HAM-AI-A-02 · Artificial IntelligenceAuditing AI Lifecycle & Data Governance ControlsEvaluate lifecycle and data governance controls across data sourcing, training, validation, deployment, monitoring, and change in ISO/IEC 42001View details →

HAM-AI-A-01 · Artificial IntelligenceAuditing AI Risk & Impact ManagementEvaluate harm, impact and risk reasoning, intended use alignment, and decision traceability in ISO/IEC 42001View details →

HAM-AI-S-03 · Artificial IntelligenceOperational Control of AI SystemsDefine, implement and maintain operational controls for AI systems across deployment, change and monitoringView details →

Advisory

We help organisations establish practical governance for artificial intelligence across strategy, risk, control, accountability and assurance. This includes structuring AI management systems, embedding oversight into the AI lifecycle, and creating documentation and evidence that support responsible use, internal governance and external scrutiny.

The four phases below are the life of a management system, and an engagement can begin at any one of them.

1. 01

   Design

   Establishing clear structures and accountability

   –AI governance framework and policy design, including AI Management Systems (AIMS) aligned with ISO/IEC 42001

   –Definition of roles, responsibilities and decision rights

   –AI system classification and risk categories

   –Integration into existing management systems (e.g. ISMS, QMS)

   –Design of documentation and evidence structures

2. 02

   Operate

   Making AI governance work in daily practice

   –AI risk and system impact assessments

   –Operational processes for AI lifecycle management

   –Controls for data quality, model changes and human oversight

   –Incident and issue handling for AI-related risks

   –Enablement of key roles (management, product owners, compliance)

3. 03

   Assure

   Providing confidence and audit readiness

   –Independent reviews of AI governance and AIMS structures

   –Control effectiveness and implementation checks

   –Outsourced internal audit based on ISO/IEC 42001

   –Certification readiness assessments

   –Supplier and third-party AI reviews

   –Preparation for internal and external audits

4. 04

   Evolve

   Keeping governance effective as technology and regulation change

   –Monitoring regulatory and technological developments

   –Scenario analysis for future AI use cases

   –Maturity assessments and improvement roadmaps for AIMS

   –Executive sparring on strategic AI decisions

   –Integration of new requirements into existing systems

Audit

An AI governance audit asks whether the oversight bites: whether the risks identified are the ones the deployed systems actually create, whether the controls hold across a change of model or supplier, and whether anyone would notice if they stopped working.