What directors in the boardroom need to know and do

June 2026

                        AUTHOR: LUIZ DE CAMPOS SALLES

                         Helper:  AI CLAUDE  SONNETT

Artificial intelligence in the boardroom: what directors need to know — and do

Boards of directors in several countries are already using AI to support their work. The question is no longer whether to engage with this technology, but how to do so responsibly. This article outlines the practical risks and concrete applications that every board should consider.

Risks that must be addressed first

Before any board considers deploying AI, two risks deserve honest attention.

First: AI systems are not infallible. They can produce confident-sounding answers that are incomplete, outdated, or simply wrong. Every output must be reviewed critically by the directors or executives who receive it. Delegating judgment to an AI — without human verification — is a governance failure, not an efficiency gain.

Second, and more consequentially: data access must be defined with precision before deployment. What information will the AI be allowed to see? Financial results? Competitive strategy? M&A discussions? The full minutes of board meetings? Each category carries a different level of sensitivity and a different risk profile.

Confidentiality is the board's primary concern. It is worth noting that a well-configured AI system may, in practice, be easier to control than the human participants in a meeting — but that does not make the configuration decision any less critical.

Boards that skip this step and grant broad data access for convenience are taking a risk they may not fully understand yet.

How AI can add real value to board work

With the right guardrails in place, AI can meaningfully improve the quality and efficiency of board work across three moments: before, during, and between meetings.

Before meetings, AI can be loaded with all materials on the agenda — financial reports, management presentations, market data — and asked to analyze, summarize, and flag issues. It can compare the company's reported results against competitors, identify trends that deserve scrutiny, and surface questions directors should be asking. Sensitive items can be excluded from this feed entirely.

During meetings, AI can serve as a real-time resource. When a director raises a question that requires data — a market size, a regulatory reference, a historical comparison — an AI assistant can provide an answer on the spot, reducing the number of questions that get deferred or go unanswered.

For strategic analysis, AI's capacity to process large volumes of data becomes particularly valuable. When a board is evaluating an international expansion, a new product line, or a major acquisition, AI can integrate market data, competitive intelligence, regulatory environments, and financial scenarios into a coherent framework to support the decision.

One especially promising application is the pre-mortem: a structured exercise in which the board assumes a proposed initiative has failed, and works backwards to identify the most likely causes. AI can accelerate and enrich this process by drawing on comparable cases, sector-specific failure patterns, and risk factors that human participants might overlook.

The bottom line for directors and CEOs

AI will not replace the judgment of an experienced board. What it can do is make that judgment better informed, more consistent, and faster to reach. The boards that treat AI as a tool to be governed — rather than a solution to be adopted uncritically — will be better positioned to capture its benefits without inheriting its risks.

The technology is ready. The governance frameworks, in most organizations, are not. Closing that gap is a board responsibility