AI has unequivocally arrived as a critical board-level concern. As we look forward to the second half of the year, AI’s influence on market dynamics, operational efficiency, and strategic direction is too profound for your board to ignore. The speed of AI development demands a new level of engagement from corporate leadership.

However, many boardrooms are inundated with technical jargon and hype, obscuring the levers that determine whether AI will drive more value or more risk. In fact, a Deloitte board governance survey found that 66% of respondents say their boards still have “limited to no knowledge or experience” with AI, 40% say AI has caused them to think differently about their boards’ makeup, and 33% are “not satisfied” or “concerned” with the amount of time their boards devote to discussing AI.

For board directors and C-suite leaders, navigating AI through the rest of the year isn’t about mastering the latest model architectures, tracking a variety of benchmarks, or deciphering the latest research. Your focus must be on understanding AI’s potential business impact, establishing robust oversight mechanisms, and ensuring organizational readiness to harness AI’s power responsibly and effectively. It’s about governing an AI-driven enterprise, not becoming AI technicians.

Why boards must engage now

Given that AI is profoundly reshaping markets and operations, transforming how businesses compete and create value, effective corporate stewardship now fundamentally encompasses a company’s AI strategy. Companies globally leverage AI for various types of automation, enhanced customer experiences, improved decision-making, and innovation across the organization, making it a cornerstone of business strategies for almost all industries, not just tech.

Failing to integrate AI can lead to lost market share, diminished brand relevance, an inability to attract talent, or even worse: a slow slide, and then sudden slip, into irrelevance. The risk of not using AI, or adopting it too slowly, is paramount, potentially impacting long-term viability and even fiduciary duties. Additionally, shareholder scrutiny regarding AI oversight is only growing, as nobody wants a stake in a company that is completely ignoring, or blindly adopting, the latest tech.

The AI market is experiencing exceptional global growth, with rapid technological breakthroughs and companies scaling AI initiatives enterprise wide. This acceleration means the window for establishing proactive AI strategies and governance is closing, and competitive advantage will stem not from mere access to AI, but from its strategic integration, robust governance, and unique application to core business problems.

All these areas demand board oversight, especially as regulatory bodies worldwide increasingly turn their focus to AI. While an AI investment pullback is inevitable, a complete pivot seems increasingly unlikely, meaning boards that don’t engage today risk being forced to play catch up tomorrow.

The four dimensions boards should focus on

To navigate AI effectively, boards should structure oversight around four dimensions: Impact, Risk, Governance, and ROI.

1. Impact: Where can AI noticeably move the needle?

Boards must guide management to focus on AI applications yielding tangible results, like cost reductions, enhanced efficiency, superior customer experiences, and new revenue streams. The true power of various AI technologies often lies in unlocking new growth strategies and business models, not just optimizing existing processes. This includes accelerating product innovation and enabling data-driven strategic pivots. Sustainable impact comes from augmenting human capabilities and transforming the workforce, requiring upskilling and a culture of human-AI collaboration. Boards must push organizations to embed AI into core processes and scale successful initiatives enterprise-wide, moving beyond “pilot purgatory.”

2. Risk: Navigating AI-specific threats

AI deployment introduces complex risks. Data privacy, security, and IP leakage are major concerns, as AI systems often use vast datasets, creating exposure to PII breaches, cyberattacks, and IP infringement if training data is misused. Algorithmic bias is another critical risk, where AI models perpetuate societal biases from training data, leading to discriminatory outcomes and reputational damage. Generative AI models can and do hallucinate, producing incorrect information that can lead to poor decisions if not managed correctly. Mitigation includes high-quality data, grounding techniques, and human oversight. The evolving regulatory landscape and the risk of “AI washing” also demand attention, with regulators scrutinizing disclosures. These risks are interconnected, meaning a holistic strategy is vital. Finally, the “black box” nature of some AI models challenges accountability, making explainable AI techniques a governance imperative.

3. Governance: Establishing clear ownership and controls

Effective AI governance requires clear structures and responsibilities. Boards must decide on oversight structures: full board, existing committees, or a new technology/AI committee. A formal AI strategic plan, aligned with business objectives and risk appetite, is essential. Adopting frameworks like NIST’s AI Risk Management Framework and ISO/IEC 42001 can provide structure while embedding ethical principles such as fairness, transparency, and accountability is crucial for trust. Governance must be agile and adaptive due to the rapid evolution of AI technologies and the less rapid, but still always changing regulations.

4. ROI: Measuring value beyond experimentation

AI initiatives must demonstrate clear ROI to justify continued investment. Organizations need a structured approach to evaluate both financial benefits (cost savings and revenue growth) and qualitative benefits (customer satisfaction and productivity). Defining measurable KPIs aligned with business objectives is crucial, covering financial metrics, efficiency gains, customer experience, and AI adoption rates. A comprehensive ROI analysis must include the Total Cost of Ownership (TCO), covering software, hardware, data, development, integration, maintenance, and talent. Boards require regular, transparent reporting on AI performance. Attributing value to AI’s enabling role in broader transformations and capturing long-term strategic benefits is a challenge, but a holistic view of AI value realization is useful. A lack of a clear ROI framework can lead to financial black holes, eroding confidence in AI programs if they fail to deliver.

Red flags around AI misalignment

Vigilance for early warning signs, including Shadow AI projects, over-indexing on vendor hype, and fragmented tech stacks, is crucial for board oversight.

“Shadow AI” refers to employees using AI tools that the IT team and leadership have not approved. Risks include data exposure (such as pasting proprietary code into ChatGPT or Claude), loss of confidentiality, reliance on misinformation, non-compliance, and security vulnerabilities. Mitigation involves clear policies, employee engagement, providing vetted tools, and robust governance. If you discover Shadow AI in your organization, it likely signals unmet business needs or slow official processes.

Another common pitfall is making AI investments based on social media hype or the fear of missing out, rather than business fit. This can lead to wasted resources on misaligned technology and introduce data security risks if vendors lack transparency. To address this type of red flag, employ rigorous due diligence on vendors, focusing on demonstrable business value.

Even if you have managed to avoid Shadow AI and are making investments based on business fit, you may end up with fragmented tech stacks that won’t scale. A patchwork of disconnected AI tools across departments can lead to complexity, redundancy, inefficiencies, and integration challenges. A unified AI stack and an AI Center of Excellence can standardize approaches and ensure alignment.

Questions every director should ask

Insufficient AI knowledge at the board level can lead to poor strategic decisions and inadequate risk oversight. Prioritizing continuous AI education for board members, potentially adjusting board composition for AI expertise, and engaging external advisors are key.

Once the board has the right knowledge, directors should ask probing questions, such as:

  • What is our enterprise-wide AI strategy, how does it align with our core business objectives, and who is explicitly accountable for its execution and outcomes?
  • How are we ensuring the quality, reliability, and ethical integrity of our AI systems, including robust data governance, bias detection and mitigation mechanisms, and appropriate levels of explainability, particularly for AI systems used in high-impact decisions?
  • What is our board-approved risk appetite for AI initiatives, and how are we proactively identifying, assessing, and mitigating key AI-related risks?
  • Are we strategically building AI capabilities into our core business processes, or are we primarily focused on peripheral pilot projects?
  • What is our defined path to enterprise-wide adoption?
  • What is our framework for measuring and reporting the ROI of our AI investments?
  • How are we addressing the critical talent, skills, and cultural shifts necessary for our organization to successfully become an AI-driven enterprise?
  • What is our comprehensive strategy for AI upskilling and reskilling?

These questions facilitate a shift from passive reception to active interrogation of AI strategy, risk, and value, ensuring intentionality in AI adoption.

Final thoughts

In June 2025, Gartner updated its Hype Cycle for Artificial Intelligence, noting that while AI investment remains strong, focus is shifting from Generative AI hype to foundational innovations like AI-ready data, AI agents, AI engineering, and ModelOps. Put another way, AI leaders are interested in strong use cases that drive ROI and avoiding AI initiatives that can’t prove their worth.

Your board doesn’t need to be made up of AI technical experts, but it must cultivate proficiency in governing an AI-driven enterprise, starting with continuous learning and asking the right questions. As organizations deploy enterprise-grade generative AI, including everything from private LLMs and secure infrastructure to continuous model tuning and domain-specific rollouts, partnering with experienced specialists is key.

AI21 Labs empowers enterprises to translate AI’s potential into tangible, sustainable business value, responsibly and effectively. The goal is to foster a culture where strategic AI adoption and robust governance enable, rather than stifle, responsible innovation, requiring a strong partnership between the board, management, and expert advisors.