AI Expertise in the Boardroom: How Advisory and Supervisory Boards Should Build It

In short: A board can only oversee AI as well as it understands it. Most boards aren’t set up for that yet. Four steps help. The board explicitly adds AI expertise to its skills matrix, appoints at least one member with operational AI responsibility, gives all members targeted training and uses AI tools in its own work. What’s needed is sound judgment. Deep technical expertise isn’t a prerequisite.

How well prepared for AI are boards today?

Poorly, but better than a year ago. In the second edition of Deloitte’s “Governance of AI” study, with 695 respondents from 56 countries, 66 percent say their board has limited or no knowledge of and experience with AI. In the previous survey, it was 79 percent.

31%

For 31 percent, AI isn’t on the agenda at all (Deloitte, Governance of AI).

In the US, according to NACD, more than 62 percent of directors at public companies surveyed say the full board now gives AI dedicated agenda time (NACD 2025).

There is no comparably robust survey for German mid-sized companies. Boards should therefore take an honest look at where they stand without waiting for benchmarks.

Why isn’t it enough to delegate AI to management?

Because a board that can’t put management’s answers in context can only approve or block. Both are risky with AI. Approve, and you may be funding pilots with no path to production or overlooking dependencies on individual model providers. Block, and you miss the window in which competitive positions are being reshuffled. Sound judgment means being able to ask questions whose answers actually decide something.

What AI expertise does a board actually need?

I distinguish three levels, and not every member needs all of them equally:

LevelWhat it coversWho needs it
Basic understandingWhat today’s AI can do and where it fails, the difference between assistants and autonomous agents, the role of data and compute costsall members
Business judgmentAssessing AI’s impact on business model, pricing, costs and competition, evaluating investments against clear criteriathe majority of the board
Operational experienceHaving introduced and owned AI in an organization, including governance, works council relations and change managementat least one member

How do you add AI to the skills matrix?

Many boards use a skills matrix covering areas such as finance, law, industry and international business. AI often shows up there under “digitalization,” a category too broad to say anything meaningful. A separate category with concrete criteria works better. Has the member owned AI initiatives? Do they know data management and regulation from their own practice? Can they judge the economics of AI applications?

An honest self-assessment of the board against these criteria is often the first finding. It shows whether a gap can be closed through training or needs to be addressed the next time a seat is filled.

How can a board build expertise without changing its members?

See the applications firsthandManagement demonstrates the most important AI applications live in the boardroom, with real data and including difficult cases.

Use the tools yourselfMembers use AI tools to prepare for meetings, for example to summarize documents or develop counterarguments, while respecting confidentiality.

Outside perspectiveOnce a year, a session with an outside guest who owns AI at another company.

Targeted trainingBoard programs with an AI focus instead of general digital seminars.

How does the board’s way of working change?

AI is also changing board work itself. When documents can be summarized in minutes and scenarios run quickly, meeting time shifts from absorbing information to discussion. That’s an opportunity, but it calls for clear rules. They set out which tools may be used with confidential documents and how results are checked. Responsibility for the judgment stays with the board.

Frequently asked questions

Does every supervisory board need an AI expert?

Not necessarily a technology expert, but at least one member with operational experience in introducing and owning AI. All other members need a basic understanding and business judgment.

How does a board measure its own AI expertise?

Through a self-assessment against concrete criteria in the skills matrix, such as AI initiatives owned, experience with data management and regulation, and the ability to judge the economics of AI investments.

May advisory board members use AI tools for confidential documents?

Only with clear rules. The board should define which tools are approved, where data is processed and which documents must not be fed into external systems.

Further reading: Five questions an advisory board should ask management about AI · Evaluating AI investments · EU AI Act for advisory boards

Sources: Deloitte (2025): Governance of AI, 2nd edition · NACD (2025): Board Practices and Oversight Survey, AI

Similar Posts