In their July-August 2025 Harvard Business Review article “How Pioneering Boards Are Using AI,” researchers Stanislav Shekshnia and Valery Yakubovich uncovered a striking paradox in corporate governance. While AI capabilities have advanced dramatically, most board members haven’t recognized their potential for board work.
The Research: A Tale of Two Boardrooms
From June to September 2024, Shekshnia and Yakubovich conducted focus groups with more than 50 board chairs, vice chairs, and committee chairs from major companies including Nestlé, Shell, Novo Nordisk, ASM, Lazard, Randstad, and Sandoz. When asked to rank various issues claiming their attention, participants placed AI relatively low—well beneath issues like CEO relationships and shareholder interactions.
Most participants reported using AI occasionally for personal needs but “never or rarely” for fulfilling their board responsibilities. As one chair told the researchers bluntly: “It has never occurred to me to consider AI for preparing for a board meeting.”
Yet buried within the same research were striking examples of directors who are using AI extensively—and experiencing dramatic improvements in their effectiveness. This contrast reveals something important about where board governance stands today.
The Pioneers and What They’re Discovering
The directors who have embraced AI for board work describe transformative results. The researchers profile “Britt,” a nonexecutive director serving on five boards in Denmark, who began using ChatGPT two years ago. She calls ChatGPT her “sparring partner” and reports that her preparation quality significantly improved while her workload decreased.
“Alexander,” a Swiss board chair, feeds his board books to ChatGPT before meetings to generate discussion questions and decision options, then engages in conversation with the tool to improve them.
According to Shekshnia and Yakubovich, these early adopters are using AI in three distinct ways:
Individual Preparation: Directors use AI to analyze management presentations, find benchmarks, formulate strategic questions, and run simulations. Rather than spending hours trying to understand what’s important in hundreds of pages of materials, they’re using AI to quickly identify key issues and strategic implications.
Scenario Planning: Boards are using AI to explore “what-if” questions that would be too complex or time-consuming for human analysis alone. The researchers describe “Gerhard,” an Austrian board chair, who used an LLM to produce scenarios about a potential acquisition target’s country. After reviewing them, the board felt the investment was above their risk appetite and rejected it. The most important outcome, however, was that management now always supports proposals with scenario analysis.
Reality Checking: Some boards are using AI to validate their strategic thinking. “Juho,” a Finnish board chair, described how after a two-day strategy retreat, he fed the prep materials to ChatGPT and asked it to select the best option. The AI’s analysis confirmed three of four conclusions, prompting deeper discussion on the fourth.
Why the Gap Exists
If AI’s potential for board preparation is this significant, why haven’t most directors discovered it? Shekshnia and Yakubovich identify several barriers:
It Simply Hasn’t Occurred to Them: Many directors think of AI primarily as an operational tool for customer service or production lines—not as something relevant to their governance responsibilities. The connection between AI capabilities and board preparation isn’t obvious until someone points it out.
Security Concerns: Directors rightly worry about feeding confidential board materials into public AI tools. As one vice chair expressed to the researchers: “To be effective, AI needs lots of data, so using it is like opening our board meetings to the public. There is no way I will allow this to happen.” This fear, while understandable, often leads to complete avoidance rather than seeking secure alternatives.
Digital Literacy: Many current directors came of age before the digital era. Learning to use AI effectively involves discomfort, potential mistakes, and unfamiliar technology—barriers that feel significant even for accomplished professionals.
Lack of Training: Even directors willing to try AI often lack guidance on how to use it effectively for board work. As one participant told the researchers: “I tried AI, but I gave up quickly. It would give me a lot of stuff, but I had no idea what was true or false. And most of it was not relevant.”
The Security Challenge
The security concerns aren’t unfounded. Shekshnia and Yakubovich acknowledge that using public AI tools like ChatGPT with confidential board materials creates genuine risks. However, they note that these risks can be managed—major AI vendors now offer enterprise versions that guarantee company data isn’t used for training, and companies like SAP are creating models trained exclusively on single clients’ proprietary data.
This points to a crucial distinction the authors emphasize: the difference between using consumer AI tools inappropriately and using purpose-built, secure AI solutions designed specifically for board governance. The security risk isn’t inherent to AI—it’s about choosing the right tools for sensitive work.
The Information Gap AI Can Close
Perhaps most importantly, the HBR research highlights something board chairs have struggled with for years: the information gap between directors and executives. Shekshnia and Yakubovich cite a 2016-2020 study in 17 European countries that identified this gap as a major challenge for boards. Chairs try to close it through site visits, subcommittees, supplementary reports, and comprehensive board books, but directors remain time-constrained and detached from operations.
This is where AI’s potential becomes clearest. As the researchers note, properly trained AI can analyze large volumes of data to discover patterns directors might miss, continuously monitor risks for early warnings, and condense information into formats that dramatically reduce processing time.
The goal, they emphasize, isn’t to eliminate director judgment—it’s to free directors from mechanical information processing so they can focus on strategic thinking. As Britt’s experience demonstrates: her level of preparation significantly improved while her workload decreased. That’s the transformation boards need.
What This Means for Board Governance
Shekshnia and Yakubovich argue that the gap between pioneering directors and the majority reveals an inflection point in board governance. A small group has discovered that AI can fundamentally improve how they prepare and engage, while most remain unaware of the possibility. This gap won’t persist indefinitely—eventually, AI-assisted preparation will become standard practice.
The question for boards isn’t whether to use AI, but how to use it appropriately. The pioneers in the study are mostly using consumer tools like ChatGPT individually and informally. While this demonstrates AI’s potential, it also creates risks and inconsistency. Some directors are well-prepared; others aren’t using AI at all. Some may be exposing confidential information; others are appropriately cautious but missing the benefits.
The Path to Purpose-Built Solutions
What boards need, according to the authors, is the middle ground: AI capabilities specifically designed for board governance, built with appropriate security from the ground up, and structured to enhance rather than complicate the board-management relationship.
This means tools that help directors understand what management presents rather than encouraging parallel analysis. It means security designed for the most sensitive information rather than retrofitted enterprise features. It means personalization based on each director’s role and responsibilities. And it means a learning curve that’s manageable for directors who didn’t grow up with AI.
The HBR research makes clear that AI’s potential for improving board effectiveness is real—the pioneers prove it. Their experiences with consumer AI tools demonstrate what’s possible. But realizing that potential across all boards requires moving beyond individual experimentation with public tools toward purpose-built solutions designed specifically for the unique needs and constraints of board governance.
The directors who haven’t yet considered AI for board preparation aren’t wrong to be cautious, Shekshnia and Yakubovich conclude. Their instincts about security risks and the need for appropriate tools are sound. What they’re missing is that AI capabilities can be packaged in ways that address those concerns—tools designed specifically for board work rather than general-purpose consumer applications.
Looking Forward
The authors end with a bold prediction: “Eventually, we believe, every board will have an AI member, maybe even one with a vote.” They argue that savvy boards will look to get ahead of this wave by becoming AI literate today. By making smarter decisions more quickly with AI’s assistance, they’ll enable their companies to move ahead of competitors—and in many cases keep that strategic advantage for some time.
The gap between pioneering directors and the majority won’t close through individual directors discovering ChatGPT. It will close when boards have access to AI tools built specifically for their governance needs—secure, purpose-built, and designed to make directors more effective within their appropriate role. That’s the future the pioneers are pointing toward, even if they’re not quite there yet themselves.
Key Challenges Identified by Shekshnia and Yakubovich:
- Most directors haven’t recognized AI’s potential for board preparation, viewing it primarily as an operational tool rather than a governance capability.
- Security concerns about using public AI tools with confidential board materials prevent many directors from experimenting with AI assistance.
- The information gap between directors and management persists despite traditional efforts, while AI offers untapped potential to close this gap effectively.
The Path Forward:
- Boards need purpose-built AI solutions designed specifically for governance—not consumer tools adapted to board use—with security appropriate for the most sensitive information.
- Directors require structured learning and support to use AI effectively, not just access to tools that leave them frustrated with irrelevant outputs.
- The focus should remain on helping directors understand management materials better, not enabling independent parallel analysis that could undermine board-management relationships.
This analysis is based on “How Pioneering Boards Are Using AI” by Stanislav Shekshnia and Valery Yakubovich, published in Harvard Business Review, July-August 2025.