Services / AI Audit & Assurance
Service
AI Audit & Assurance
Independent, evidence-based assurance that your AI systems work as intended, treat people fairly, stay secure and comply with the law.
Overview
Assurance Boards and Regulators Can Rely On
Boards, regulators and customers increasingly ask one question about AI: how do you know it works? GCATI audits AI systems across their lifecycle, from data and model development to deployment, monitoring and retirement.
We apply audit-grade methodology to new AI risks such as bias, hallucination, prompt injection and autonomous agent behavior.
Best Suited For
- Internal audit teams adding AI to the audit plan
- Organizations deploying high-risk AI
- Firms subject to the EU AI Act, NYC Local Law 144 or Colorado’s AI Act
- Audit committees seeking AI assurance
The Challenge
Why AI Needs Its Own Audit Approach
New Risks, Old Audit Programs
Traditional IT audit programs were not built for model drift, bias or generative AI behavior.
Black-Box Systems
Vendor and foundation models can be hard to inspect, making evidence difficult to gather.
Required Audits and Assessments
Some laws now require bias audits, conformity assessments or impact assessments for specific AI uses.
Limited In-House Expertise
Few audit teams combine audit methodology with a working knowledge of AI technology.
What We Deliver
A Complete AI Audit Program
AI Audit Universe and Plan
A risk-based inventory of auditable AI and a multi-year audit plan.
AI Governance Audits
Testing of AI policies, roles, committees, inventories and approval processes.
AI Lifecycle Control Testing
Data, development, validation, deployment, change and monitoring controls.
Bias and Fairness Testing
Outcome analysis across protected groups and review of mitigation measures.
Explainability Reviews
Whether decisions can be explained to customers, regulators and staff.
Generative AI and Agent Security
Prompt injection, data leakage, excessive agency and guardrail testing.
Privacy Assessments for AI
Lawful data use, consent, retention and automated decision rights.
Regulatory Readiness Assessments
Gap analysis against the AI laws and standards that apply to you.
Standards and Regulations
Audit Criteria We Use
NIST AI RMF
Govern, Map, Measure, Manage
ISO/IEC 42001
AI management system
EU AI Act
Risk tiers and obligations
NYC Local Law 144
Bias audits for hiring tools
Colorado AI Act
High-risk AI decisions
SR 26-2
Model risk management
OWASP Top 10 for LLM Applications
Generative AI security
IIA Global Internal Audit Standards
Internal audit practice
Our Approach
Plan, Test, Report, Validate
Plan
AI inventory review, risk assessment and agreed audit scope.
Output: Audit plan and test program
Fieldwork
Walkthroughs, control testing, model and outcome analysis.
Output: Documented evidence and draft findings
Report
Rated findings with root causes and practical recommendations.
Output: Audit report and management action plans
Follow-Up
Independent validation that issues are fixed before closure.
Output: Issue validation and closure memos
Outcomes
What You Gain
- Independent confidence that AI works as intended
- Evidence for regulators and audit committees
- Early detection of bias, drift and security gaps
- Stronger AI controls through practical recommendations
- Clear ownership of remediation
- An AI audit program your team can sustain
FAQ
Common Questions
Can you audit AI we buy from vendors?
Yes. We assess vendor documentation, contracts and testing evidence, along with your own controls over how the AI is used.
Do you perform bias audits for NYC Local Law 144?
We can perform independent bias assessments and help you meet the law’s audit and notice requirements.
Can you work with our internal audit team?
Yes. We can lead the audit, co-source with your team or provide subject matter experts.
How is an AI audit different from an IT audit?
An AI audit adds model performance, data quality, fairness, explainability and AI-specific security to traditional IT control testing.
Ready to Govern and Scale Your AI?
Start with a complimentary AI risk and readiness conversation.