AI Governance Assessment & Assurance
Productised, evidence-led assessment of AI systems. We establish what applies, what evidence exists, where the gaps are and what an authorised human reviewer can reasonably conclude.
Services
Three ways Eshalu can help.
Independent AI governance assessment, responsible AI consulting and practical training — scoped separately so the role Eshalu plays is always clear.
Discuss your AI systemsView engagement models
What we do
Assess. Advise. Train.
The three service lines can be bought separately. The key rule is simple: Eshalu does not independently assess the same AI system or control work that it has advised on, designed or helped implement.
AI Consulting & Responsible Implementation
Advice and hands-on support across AI strategy, use cases, value, solution design, testing, governance, data and responsible implementation.
AI Training & Capability Building
Role-based learning for boards, executives, governance teams, business users, product teams, technical teams and data teams.
01 — Assessment & assurance
Independent assessment, backed by evidence
Eshalu's productised assessment creates one governed record per AI system and separates organisation-wide governance from the system-specific evidence that supports each conclusion.
System and governance profile
Purpose, owner, supplier, lifecycle, autonomy, impact, responsibilities and regulatory signals for the individual AI system.
Risk and control assessment
Governance, accountability, human oversight, security, third-party risk, testing, monitoring, change and operational controls assessed in context.
Data governance assessment
Evidence around data provenance, permitted use, quality, sufficiency, representativeness, lineage, bias, documentation and ongoing monitoring.
Evidence record
What was requested, supplied, accepted, missing, reused or due for refresh, with version and custody information.
Reviewer-confirmed findings
An authorised human reviewer challenges the evidence and confirms what the record does — and does not — support.
Reports and action requirements
Audience-specific outputs, prioritised findings and the evidence or control outcomes required to address each gap. Eshalu may explain the finding without designing the client's solution.
02 — AI consulting
From AI opportunity to responsible implementation
For organisations that want help deciding where AI can create value, shaping the solution and putting the right governance, data and testing foundations around it.
AI strategy and operating model
Business objectives, AI ambition, operating model, governance structure, decision rights, roadmap, investment priorities and build-versus-buy choices.
Use cases, value and benefits
Identify, challenge and prioritise AI use cases; define expected benefits, costs, dependencies, risks, feasibility and measurable success criteria.
Solution and responsible-design strategy
Architecture and delivery choices, human oversight, autonomy boundaries, supplier considerations, control-by-design and implementation governance.
Testing and evaluation strategy
Define what good looks like before deployment: evaluation criteria, acceptance thresholds, scenario coverage, safety testing, robustness, human review and post-deployment monitoring.
Lawful and traceable data acquisition
Data source strategy, provenance and lineage, rights and licensing, permitted use, privacy considerations, collection controls and documentation needed to show where data came from and why it may be used.
AI data quality and fitness
Quality, completeness, sufficiency, relevance, timeliness, representativeness, bias and imbalance, training/validation/test separation, synthetic data and monitoring for drift or changing suitability.
Independence boundary: consulting can include design and implementation support. If Eshalu materially advises on, designs or helps implement a system or control, Eshalu will not later present itself as the independent assessor of that same work.
03 — Training
Build the capability to use and govern AI well
Practical, role-based learning rather than generic AI awareness. Training can be delivered as executive briefings, team workshops or tailored capability programmes.
Boards & executives
AI opportunities, accountability, investment decisions, governance, risk, regulatory expectations and the questions leaders should be asking.
Governance, risk & legal teams
AI governance operating models, evidence, controls, risk assessment, regulatory concepts, standards and assurance readiness.
Business, product & delivery teams
Finding useful use cases, defining benefits, responsible adoption, human oversight, supplier use and what good delivery evidence looks like.
Technology & data teams
Responsible design, testing and evaluation, data provenance and quality, monitoring, change control and evidence that supports governance decisions.
Clear boundaries
Helpful where we advise. Independent where we assess.
Assessment
We assess governance and evidence, record gaps and provide human-reviewed conclusions. We can explain required outcomes, but we do not design or implement the controls we later assess.
Consulting
We can advise on strategy, design, data, testing, governance and implementation. That work is clearly identified as advisory and is excluded from Eshalu independent assessment.
Training
We build capability without taking accountability away from the organisation. The client remains responsible for its decisions, policies, systems and legal advice.