Directors at mid-market companies have reached a threshold. Nearly every one of them now talks about artificial intelligence in the boardroom. A survey released today shows 98% of such firms have discussed involving AI in high-level decisions. The findings arrive as larger enterprises install chief AI officers and global boards slowly add the topic to standing agendas.
Board Intelligence, a London-based provider of board software and advisory services, polled just over 400 directors, CEOs and CFOs in the UK, US, Scandinavia and Middle East. Respondents came from companies with annual revenues between roughly $67 million and $668 million. The results paint a picture of urgency mixed with caution. Forty-nine percent have moved beyond talk into implementation, actively deciding which board processes to hand over to AI systems. Another 34% discussed the technology at board level without specific proposals. Fifteen percent left the matter to management.
Only 37% of those surveyed called their boards an essential tool for value creation. The dissatisfaction runs deeper. Eighty-six percent reported that overly rigid or inconsistent processes produced poor, delayed or rushed decisions in the past six months. Forty-one percent said boards waste too much time on past performance reviews. And while 79% believe their boards enable innovation, just 18% said they do so strongly.
Megan Pantelides, senior director at Board Intelligence, pointed to familiar boardroom headaches. “The problems AI can realistically address include information overload, information asymmetry versus management, inadequate preparation time and cognitive bias,” she told TechRepublic. AI, she added, performs the analytical heavy lifting. It synthesizes large volumes of information, surfaces patterns, flags risks and assumptions buried inside dense board packs.
Yet directors draw firm lines. Legal accountability still rests with humans. Pantelides noted that boards recognize limits. “The fact that such a high proportion of boards are discussing which decisions should remain human-led indicates that boards recognise there are limits to AI’s role in governance.” Decisions on policy changes might suit AI support. Calls on mergers and acquisitions, culture, leadership, strategy or crisis response do not. “In the main, boards are not prepared to delegate the latter — and nor should they.”
This tension between analysis and judgment echoes across recent research. Four out of five boards now actively consider where human judgment ends and AI begins, according to the same Board Intelligence work covered days ago by Board Agenda. Pippa Begg, CEO and co-founder of Board Intelligence, framed the moment sharply. “Boards are being forced to confront a fundamental question: where should human judgement end, and where should AI begin?” She warned that many governance structures lag the speed and complexity of technological change, especially when overseeing AI adoption itself.
Progress shows elsewhere. Deloitte’s second global survey on AI governance, conducted in January and February 2025 with 695 board directors and executives across 56 countries, found only 31% still keep AI off the board agenda. That figure dropped from 45% in the prior year’s poll of 468 respondents. Boards reporting limited or no AI knowledge fell from 79% to 66%. Forty percent now say AI influences board composition, up four percentage points. Still, one-third of directors remain unsatisfied with the time spent discussing the technology.
Larger organizations reveal sharper divides between success and experimentation. Protiviti’s 2026 Global Board Governance Survey reported that just 26% of boards discuss AI at every meeting. Among companies seeing strong returns from AI initiatives, that share jumps to 63%. For those with weaker results, it sits at 13%. Ninety-five percent of boards confident in their ability to integrate AI report significant ROI. Only 33% of less confident boards can say the same. “Success is defined not by having all the right answers, but rather by asking the right questions,” the report observed.
CEOs have taken notice. IBM’s 2026 CEO study found 76% of organizations now have a chief AI officer, up from 26% the year before. CEOs expect the share of operational decisions made by AI without human intervention to rise from 25% today to 48% by 2030. BCG’s AI Radar survey showed nearly three-quarters of CEOs position themselves as the primary decision-maker on AI strategy, double the share from the previous year. Half believe their job stability hinges on getting AI right this year.
Yet adoption among executives themselves remains modest. Research highlighted by Time in March, drawing on work by Stanford economist Nick Bloom and collaborators, found 69% of CEOs, CFOs and senior leaders use AI at work less than one hour per week. Twenty-eight percent use it not at all. PwC’s 29th Global CEO Survey released in January reported that while 30% of CEOs saw revenue gains from AI in the past year and 26% saw cost reductions, 56% reported neither benefit.
Real-world experiments remain rare and carefully bounded. Mubadala, the Abu Dhabi sovereign wealth fund, placed an AI member on its investment committee. Lloyds Banking Group began using Board Intelligence tools in April to analyze and summarize documents ahead of meetings. Future applications could include bias detection in decision-making or support on cybersecurity, sustainability and market analysis. Directors still review and authorize outputs.
Academics and lawyers urge structured oversight. Floris Martens of the University of Ghent suggested that directors’ duty to act on an informed basis may evolve into a duty to rely on AI output when its analytical capabilities surpass human ones. Researchers Maria Lucia Passador and Maria Lillà Montagnani wrote that Europe’s business judgment rule should protect leaders only if they engage critically with algorithmic tools, demand traceability and document their rationale for using machine-generated insights.
Protiviti’s data underscores the payoff. Boards that treat AI as a standing agenda item gain visibility into opportunities and risks. They build confidence in responsible integration. Organizations with mature AI practices cite improved customer experience and process efficiency more often than laggards. Innovation risk tops concerns for advanced adopters, followed by talent gaps and infrastructure limits.
The pattern repeats. Talk has become action for a small but growing group. Most boards still sit in the middle, discussing without deep integration. Mid-market directors, squeezed by information overload and part-time service across multiple boards, appear furthest along the curve. They see AI as a way to challenge management more effectively and reach decisions with greater confidence. But the final call stays human. Accountability demands it.
And the gap between awareness and structured governance remains wide. Only a minority of Fortune 100 companies disclose formal board-level AI oversight. Fewer still have written policies approved at board level. Directors who master the balance between machine analysis and human judgment will likely separate winners from also-rans in the years ahead. The data says most understand the stakes. Implementation, as always, proves harder.
WebProNews is an iEntry Publication