Artificial intelligence (AI) is rapidly transforming internal audits. From detecting transaction anomalies to enabling continuous monitoring, AI promises faster insights and broader coverage than traditional audit approaches. But there’s a problem: many stakeholders don’t fully understand how AI works or why they should trust it.
If your board, executives or business leaders feel like AI-driven audits are a “black box,” you’re not alone.
The good news? COBIT offers a practical solution through “EDM05: Ensured Stakeholder Engagement.” When applied correctly, EDM05 helps bridge the gap between complex AI tools and the people who rely on their outputs.
Here’s how to make AI-powered audits more understandable and more trusted.
Start With Expectations, Not Technology
One of the biggest mistakes organizations make is leading with technology.
Instead, start with conversations:
- What risks matter most to your stakeholders?
- What decisions are they struggling to make?
- What concerns do they have about AI?
Then translate your AI capabilities into plain business language.
- Instead of saying, “We’re deploying an anomaly detection model…”
- Say, “We’re using AI to scan all transactions and flag unusual activity for review.”
This simple shift makes AI feel like a tool rather than a mystery.
Communicate Like a Business Leader, Not a Data Scientist
Different stakeholders need different levels of detail.
- Board/Audit Committee: Focus on risk, trends, and outcomes
- Executives (CFO/CIO): Emphasize business impact and operational insights
- Audit Team: Provide deeper technical detail and documentation
The key is consistency: always connect AI outputs to business VALUE.
For example:
- “Detected $2.3M in unusual transactions” is good.
- “Identified $2.3M in high-risk transactions, reducing potential fraud exposure” is better.
Also, avoid overloading stakeholders with technical jargon. If they can’t explain it back to you, it’s probably too complex.
Make AI Explainable—Not Just Accurate
Accuracy alone does not build trust; explainability does.
When AI flags an issue, stakeholders should understand why:
- Was the transaction outside normal ranges?
- Did it bypass approval controls?
- Does it deviate from historical patterns?
Clear explanations turn AI from a black box into a decision-support tool.
Build Trust Through Transparency
Trust isn’t automatic; it’s earned. Here are three practical ways to build it:
-
Be Honest About Limitations
- AI isn’t perfect. It produces false positives and depends on data quality. Say that upfront.
-
Show Governance in Action
Stakeholders should know:
- Who owns the AI models
- How performance is monitored
- How changes are approved
This reinforces that AI is controlled and not operating unchecked.
-
Demonstrate Fairness
- Bias testing isn’t just a technical exercise; it’s a trust-building activity.
- Share results in simple terms: “We tested the model across regions and found consistent accuracy rates.”
- Transparency builds credibility.
Use Agile to Keep Stakeholders Engaged
If you’re using Agile or Scrum in your audit process, you already have a built-in advantage.
Leverage:
- Sprint Planning to align priorities
- Sprint Reviews to demonstrate findings
- Retrospectives to gather feedback
This creates a continuous feedback loop where stakeholders aren’t just informed, they’re involved.
Close the Loop: Show That Feedback Matters
Nothing builds trust faster than action.
If stakeholders say:
- “The dashboard is too complex.” → Simplify it.
- “We don’t understand the metrics” → Refine the visuals.
- “There are too many false positives.” → Adjust the model.
Then communicate the improvements back to them.
When stakeholders see their input reflected in changes, engagement increases, and so does confidence.
Measure What Matters
To improve stakeholder engagement, track it. Remember that if you can’t measure it, you can’t manage it.
Consider measuring:
- Stakeholder satisfaction with audit reporting
- Understanding of AI outputs
- Confidence in AI-driven findings
If stakeholders don’t trust or understand the results, even the best AI models won’t deliver value.
Final Thought: AI Success Depends on People, Not Just Technology
AI can transform internal audits, but only if stakeholders trust it.
COBIT’s EDM05 reminds us that governance is not just about systems; it’s about people, communication and alignment.
If you focus on clear expectations, simple communication, transparent practices and continuous feedback, you can turn AI-powered auditing from a black box into a trusted strategic asset.
If you’re implementing AI in your audit function, don’t just ask, “Is the model working?”
Ask, “Do our stakeholders understand and trust it?”
That’s where the real business VALUE begins!
