71% of directors say their boards need stronger AI skills, yet only 27% call specialized expertise such as AI very important in candidates. The profile comes first, and I'd prioritize an operator who has deployed AI.
Both figures come from PwC’s 2026 survey, in which 81% of directors rated alignment with the board’s culture as very important when evaluating prospective directors. Three times as many gave that rating to fit as to specialized expertise, which raises a fair question about whether nominating criteria match the capabilities boards say they lack. Cisco offered a recent example on October 9, when it added Girish Rishi to its board as an experienced technology operator with explicit data and AI credentials. Rishi runs the industrial software company Cognite as chairman and CEO.
Why Do Boards Lack AI Expertise When Directors Say They Need It?
ISS-Corporate’s review of disclosures through 2025 found that 69% of S&P 500 companies had no directors with clearly disclosed AI skills. Morgan Stanley’s August 2026 analysis of ISS data compiled in June found at least one AI-skilled non-executive director on 68% of S&P 500 boards, using a broad definition that includes prior experience in machine learning or a related topic. The ISS-Corporate and Morgan Stanley figures draw on different periods and definitions, so they aren’t directly comparable. Either way, PwC found that 82% of directors rate the information they get on AI’s link to risk and business performance as fair or worse. Without a director who can read a model evaluation or question a deployment timeline, a board has little basis for asking management to report better.
What Does AI Regulation Change for Board Recruiting?
Under the EU’s AI Omnibus, adopted by the Council on 29 June 2026, the AI Act’s requirements for high-risk systems listed in Annex III generally apply from 2 December 2027, with qualifying product-related systems following on 2 August 2028. Annex III includes systems used in recruitment and selection and those that assess people’s creditworthiness, subject to classification rules, exceptions and transitional provisions. Other obligations, such as the rules on prohibited practices, already apply.
Director searches can take months, and a new board member may need time to settle into committee work, so someone recruited in mid-2027 could still be learning the business when the high-risk requirements take effect. Boards expecting some of their systems to be classed as high-risk have reason to start this year. Day-to-day AI governance may sit with a Head of AI Safety and Governance, while the board oversees management's approach.
Which AI Director Profile Does Your Board Need?
AI expertise covers at least four backgrounds in the framework below, which is a starting point and has overlaps, since a former Chief AI Officer can count as both an operator and a technical specialist.
[Table 1]
Paul DeNicola of PwC’s Governance Insights Center told Bloomberg Law that AI expertise has no single definition, and he noted that some board consultants consider stacking boards with issue-specific directors a mistake. I’d choose the operator at most companies, unless technical or regulatory exposure calls for a specialist.
A director seated only for AI risks being asked only about AI.
The decisions that shape a company’s AI direction often come up in capital allocation and CEO succession discussions, and an operator brings deployment experience into those debates. Companies that train their own models or sell AI as the product are the exception. Their boards need a technical specialist, ideally seated alongside an operator.
Which AI Decisions Will This Director Weigh In On?
Before the search opens, the nominating committee should be able to name the AI decisions the new director will help oversee in their first year. Examples include a build-or-buy choice on models, a high-risk system heading toward the EU deadline, or an acquisition with AI at its center. A committee that can’t name one is still working out what the role is for.
When Does a Board Need a New Chair or an Advisory Board?
A new chair makes sense once AI changes what the company sells or how it competes. The chair sets the agenda and leads board refreshes, and PwC’s survey shows appetite for refresh: among the 55% of directors who think a colleague should be replaced, 39% cite insufficient expertise, up from 21% in 2025. Candidates for a technology chair can come from former CEOs who took a company through an IPO or sale, and from people already chairing other technology boards. Bexorg, which builds neurology drugs on an AI brain platform, took a domain route on October 27 when it named Mikael Dolsten chairman, bringing large-pharma R&D leadership to the role.
An advisory board answers a narrower need, because advisers generally have no vote and don’t owe directors’ fiduciary duties just by holding the role, though their obligations depend on jurisdiction, contract and conduct. That flexibility suits sitting executives who can’t join another board. On October 27, Potentially AI named technology lawyer Josje Koehof Adviser to the Board on AI Governance and Data Strategy, with a brief covering its pledge not to train models on user data. An adviser suits a defined technical problem management wants help with, and the board seat remains the place to challenge the decisions management brings forward. In my view, companies get more from two advisers with a clear mandate and regular time with the CEO than from a long list of well-known names, and a written commitment makes the ask easier for senior people to accept.
Where Do Strong AI Director Candidates Come From?
I’d look first at sitting executives at companies with AI in production, with titles such as Chief AI Officer, CTO, VP of AI, Head of AI and SVP of Engineering. Plenty of them have never held a public board seat, so boards open to first-time directors widen their options, though employer policies often limit outside seats.
How Should Boards Assess AI Fluency in a Director Candidate?
[Table 2]
References from former direct reports are worth seeking alongside those from CEOs, because the people who worked under a candidate usually know whether that person understood the technology or relied on the team to explain it.
What Should Companies Look for in an AI Board Search Partner?
Recent board appointments of directors with AI or machine learning backgrounds, and whether those directors are still serving, say more about a firm than its pitch. Ask how it approaches sitting executives who aren’t looking for a board seat, since they make up much of the AI director pool. Off-limits agreements belong in the same conversation, because the firm’s list shows which parts of the candidate pool it can and can’t approach. It’s also worth asking how the firm tests candidates’ deployment experience and board judgment, since an executive title says little about either. Confidentiality matters most when the search could lead to replacing a sitting director.
Where This Approach Falls Short
A company using vendor AI in back-office work may get enough from director education plus outside advisers, both routes PwC lists alongside new appointments. Alston & Bird, writing for Law360, notes that education works best when at least one director already has a grounding in the subject. Operators with AI deployment records are in heavy demand, and cultural fit still counts for something, since a director who can’t work with colleagues will struggle to change much on any board.
Frequently Asked Questions
How Many Directors With AI Expertise Should a Board Have?
There’s no established benchmark. Morgan Stanley’s analysis found 24% of S&P 500 boards with three or more AI-skilled non-executive directors. For companies building AI products, I’d consider pairing a technical specialist with a director who has scaled a business.
Does the EU AI Act Require Boards to Have AI Expertise?
The Act sets obligations for companies and doesn’t prescribe board composition. Boards still oversee compliance, and Annex III high-risk requirements generally apply from 2 December 2027.
What Is the Difference Between an AI Advisory Board Member and a Non-Executive Director?
A non-executive director votes and shares responsibility for holding management to account. Advisory board members give counsel on defined questions, generally without a vote or directors’ fiduciary duties.
How Should a Board Run an AI Director Search?
- Name the AI decisions the director will oversee, then choose the profile and committee seat to match.
- Give AI expertise a weight in the skills matrix, scored next to cultural fit at shortlist stage.
- Map candidates by the systems they’ve put into production, since titles such as Head of AI cover very different jobs from one company to the next.
- Check outside-board limits early, because employer approval can add weeks the board didn’t plan for.
- Plan onboarding before the vote, including a briefing from the company’s AI leadership and access to its AI inventory.
Boards preparing that kind of search can contact Christian & Timbers to discuss a confidential director or chair search.
