AI ASSURANCE
Near-evaluation of AI governance structures, risks & controls, and individual AI application cases
Independent evaluation of your AI deployment
AI Assurance provides an independent view of the suitability of AI governance, controls, and evidence for controllable and verifiable AI deployment.
Internal and external stakeholders want to be able to trust that management is aware of and has control over the AI-related control, liability, and reputation risks.
The focus is on four performance areas:
Governance Assessments · Risk & Control Assessments · Review and Monitoring of AI-related Documentation · AI-related Audit and Validation Services
I offer these AI Assurance services to you
AI Governance
Assessment of Objectives, roles, responsibilities, decision paths, guidelines, approvals, monitoring, and governance documentation
AI risks & controls
AI-based risk analysis, implementation or review of the adequacy and effectiveness of selected controls
AI-related documentation (review & monitoring)
Completeness, consistency and timeliness of AI inventory, risk analyses, releases, control evidence and change documentation
Confirmation services
Examination of specific AI-related issues for internal or external stakeholders based on agreed criteria
From focused assessment to performance validation
The scope of the examination depends on the question being asked.
Typical question
How far are we from our goal? What are the main gaps?
Scope of delivery
Limited assessment of selected governance structures, risks, controls, or existing documentation
Result
Status picture, essential findings, recommended areas of action
Typical question
Is the documentation provided still current and appropriate?
Scope of delivery
Focused and, if necessary, recurring assessment of agreed facts
Result
Key findings, observed changes, identified action needs
Typical question
Are the defined controls actually implemented, effective, and appropriate?
Scope of delivery
Risk-oriented testing procedures based on agreed criteria and evidence
Result
Examination findings and evaluation
Typical question
Do internal or external stakeholders need a formal independent statement regarding AI governance structures, processes, controls, or documentation?
Scope of delivery
Testing according to appropriate criteria and agreed-upon AI-related laws and standards (e.g., AI-VO, ISO/IEC 42001:2023, NIST)
Result
Formal statement with explanations
AI Assurance in practice
From the review of AI-based financial processes to the assessment of governance across departments or companies: two examples show possible audit topics, procedures, and results.
The degree of reliability of the tests in confirmatory procedures
Provided that the subject matter, criteria, and evidence base allow for a formal confirmation performance, the contract can be structured with varying levels of assurance.
Limited security
The testing procedures are planned and conducted in such a way that a meaningful but limited level of testing certainty is achieved.
Adequate security
The testing procedures are planned and carried out in such a way that an increased, but not absolute, level of testing reliability is achieved.
The entry questions for each confirmation task
Regardless of the scope of the examination, the assessment of the subject matter begins with these questions:
1. What is essential for the test statement?
2. What risks exist for the reliability of the statement?
3. By what evidence can the examination statement be substantiated?
Clearly defined scope of services instead of an open consulting project
The subject matter, scope of work, criteria, reporting, and the intended level of assurance will be determined transparently before the start.
For clearly standardized assessments and reviews, especially monitoring, fixed performance scopes and prices can be agreed upon.