AI is moving from a niche technology topic to a core business issue, and boards are being asked to oversee its impact on business models, productivity, security, and talent. Research from KPMG International and the INSEAD Corporate Governance Centre shows that AI will fundamentally reshape how organizations operate and how boards set priorities.

The challenge is that most boards are not yet ready for this shift. The report highlights that nearly 75% of boards still have only moderate or limited AI expertise. That means many directors struggle to:

• Understand how AI and human decision-making should work together.
• Assess AI-related risks and opportunities without over-relying on management.
• Provide credible oversight of AI strategy, ethics, and compliance.

As AI becomes embedded in core processes and customer experiences, boards can no longer treat it as a purely technical or IT issue. It is now a strategic governance responsibility that touches risk, culture, workforce, and long-term value creation.

Boards can move from abstract discussions about AI to practical oversight by focusing on three areas: director education, board structure, and governance processes.

1. Elevate AI fluency through hands-on learning

• Go beyond high-level briefings and mandate experiential learning—for example, workshops where directors actually use generative AI and agentic tools to see their capabilities and limitations.
• Encourage individual directors to practice with AI themselves, attend AI-focused programs for board members, and visit companies that are advanced in AI implementation.
• Consider an external AI advisory council or independent technology advisors to provide objective perspectives that are distinct from management’s view.

2. Adjust board and committee responsibilities

• Make AI strategy and risk a standing agenda item at the full board level, not just an occasional topic.
• Update committee charters so AI is explicitly covered:
  • Audit committee: AI-related assurance and controls.
  • Risk committee: AI’s impact on enterprise risk.
  • Compensation committee: Linking incentives to responsible AI adoption, ROI, and workforce transformation metrics.
  • Nominating/governance committee: Tracking and improving board tech and AI fluency.
• Some organizations create a dedicated AI or technology committee, but the main board still needs time on the agenda to review its reports and recommendations.

3. Embed trustworthy AI into governance

• Set expectations that the organization will build, buy, and use AI in ways that are lawful, ethical, robust, and aligned with enterprise values.
• Ensure there are documented processes that translate ethical guidelines into practice—from AI design and sourcing through to ongoing monitoring.
• Oversee data governance so AI systems use data that is obtained legally and protected from misuse or unauthorized access.
• Require management to maintain a clear, defensible approach to regulatory and legal obligations related to AI.

By combining education, structural changes, and clear guardrails, boards can move quickly from awareness to meaningful AI oversight.

Boards are being asked to help the organization reimagine how AI is used—without losing sight of risk, accountability, and people. Three themes are emerging: human accountability, controlled experimentation, and workforce transformation.

1. Keep humans accountable for high-impact decisions

• Use AI to augment decisions, not make them outright, especially for high-impact areas.
• For critical decisions, insist on meaningful human oversight where AI outputs are explainable, auditable, and aligned with the company’s values, risk appetite, and regulatory requirements.
• Ask management to clearly define how AI-driven and human-driven decision-making interact, and then evaluate whether that approach is thorough and thoughtful.

2. Encourage innovation in controlled environments

• Support a secure “sandbox” approach—rapid experimentation within ring-fenced environments where data privacy and security are protected.
• Approve agile governance models that:
  • Fast-track low-risk, internal AI use cases.
  • Require more rigorous review for high-risk, customer-facing applications.
• Expect management to use a disciplined innovation process with clear rules, priorities, and criteria for scaling or stopping AI initiatives.

3. Treat AI as a workforce transformation, not just cost-cutting

• Set the tone that AI should be a human augmentation tool, not simply a headcount reduction lever.
• Require a comprehensive workforce transformation strategy that covers upskilling, reskilling, and redeploying employees to build a sustainable human–AI environment.
• Ask management to articulate how specific tasks and processes may evolve, what skills will be needed, and over what time frames.
• Given uncertainty, expect investment in broad AI skills training so the organization can stay flexible and pivot as technology and regulations change.

By focusing on these areas, boards can help the organization capture AI’s benefits while maintaining trust, protecting the enterprise, and supporting long-term talent and culture.