Most leadership teams are making bigger decisions faster than at any point in their careers. Almost none of them have fundamentally changed how they make decisions – and this is the gap AI is exposing — and it is getting worse each day.
The consistent themes are emerging in 2026 research:
- Leaders are not ready for AI deployment that is already happening in their organisations.
- AI generated decisions are being acted upon without any governance policies in place or at best, they are operating with generic AI frameworks.
- Few employees in AI transformations feel the scope and rationale for change was clearly communicated.
- AI pilot failure is not technological it is human and primarily it is leadership-level governance that is not keeping pace.
- AI cannot replicate distinctly human dimensions of leadership: judgement, empathy, purpose and accountability.
- Leadership teams now require a new kind of coaching beyond a periodic personal intervention to one that is embedded in the context of decision-making, governance structures and real operating conditions. And they need to begin this new process fast.
THE SHIFTING GROUND
The assumptions that underpinned leadership are under pressure
For decades, effective leadership rested on a familiar premise: better information leads to better decisions, which leads to better alignment of people. Superior analysis was itself a source of competitive advantage. The executive who commanded the most complete picture of a situation was best placed to act.
That premise has not disappeared. But the environment in which it operates has changed profoundly. AI tools can now synthesise information, test scenarios, challenge assumptions and generate strategic options in seconds. Capabilities that once required teams of analysts, weeks of data gathering, or expensive external advisors are increasingly available on demand. McKinsey’s 2026 Global Tech Agenda documents leading companies “investing heavily to scale agentic AI systems that autonomously plan, decide, and act across workflows” — systems capable of reshaping how; and the speed that decisions get made.
Inside many organisations today, this means that decisions which once moved through weekly management rhythms are now happening continuously, through AI-enabled workflows and automated recommendations. The operating environment is accelerating. Leadership structures, in most cases, are not.
“The organisations that thrive will be those that move from reacting to technology to shaping it — regularly stress-testing their operating models, leadership roles, and talent pipelines against alternative futures.” — McKinsey, State of Organizations 2026
This is not the end of the leadership approaches we have known. But it is the beginning of a very different leadership reality — one that the evidence suggests most organisations are not yet equipped to navigate.
THE EVIDENCE
2026 research reveals an uncomfortable gap
Deloitte’s 2026 Global Human Capital Trends research, conducted with Oxford Economics across 89 countries, sharpens the picture further. Sixty percent of executives report using AI in decision-making. Just 5% say they manage it well. Deloitte identifies this as the accumulation of what they call ‘culture debt’ — the cost organisations incur when they scale AI without building the accountability structures and trust frameworks to govern it.
McKinsey’s State of Organizations 2026, drawing on more than 10,000 senior executives across 15 countries, found that 72% of leaders describe their organisations as not fully ready for the changes already underway. Notably, this is not a fringe finding among laggards. It is the majority view, including among leaders who remain broadly optimistic about their direction.
5% of Executives using AI in decision-making say they manage it well — despite 60% reporting active use. (Deloitte Global Human Capital Trends 2026)
BCG’s 2026 research across 2,400 executives, including 640 CEOs, reveals a further dimension of the challenge. While nearly three quarters of chief executives now identify as their organisation’s primary decision-maker on AI — double the proportion from 2025 — only 15% qualify as what BCG terms ‘trailblazers’: leaders who are systematically upskilling their workforce and building reinforcing cycles of adoption and confidence. The remaining 85% are engaged but not yet executing at the level the moment requires.
For mid-market businesses — those typically running between 50 and 500 people — these dynamics are arguably more acute. Research from Everest Group in early 2026 found that only 7% of mid-market enterprises have put agentic-specific governance policies in place, with approximately 30% operating with either generic AI frameworks or no policy at all. The governance gap is real and will create real consequences unless addressed.
7% of mid-market enterprises have created agentic-specific governance policies (Everest Group Research, 2026)
THE CORE ARGUMENT
The gap is not about information — it is about judgement!
Most organisations do not lack information. They have more data, dashboards and AI-generated insight than their leadership teams can comfortably absorb. The challenge is something different and more fundamental: how do leaders make sound decisions when speed, ambiguity and technological change are all increasing simultaneously?
AI is highly effective at analysis, synthesis and pattern recognition. It can surface options, expose assumptions and identify signals that humans would otherwise miss. Bain’s research on AI operating models is unambiguous on this point: leaders are “no longer just coordinating who decides what and when — they are designing systems in which high-quality decisions can be made quickly and consistently across the organisation.”
But AI cannot own the decisions it informs.
AI cannot will not be able to effectively carry out the following:
- Exercise political judgement
- Read organisational culture
- Maintain human trust under conditions of genuine uncertainty.
As Harvard Business Review noted in early 2026, drawing on research by Martin Reeves and colleagues at BCG Henderson Institute, “many crucial aspects of decision-making lie beyond” what automated systems can replicate — specifically those involving contextual awareness, ethical sensitivity, and the interpretation of weak and ambiguous signals.
“AI accelerates us. Judgement guides us.” — The defining leadership principle of high-performing organisations in 2026
Deloitte’s 2026 Human Capital Trends research frames this as an imperative to treat decision-making as a “strategic discipline” — intentionally designing how humans and AI share judgement and accountability, rather than allowing that boundary to drift by default. Few organisations are doing this. Most are allowing the boundary to be set by the technology rather than by the leadership team.
THE LEADERSHIP IMPLICATION
The traditional model of developing leaders is no longer sufficient
This is where the most significant — and least discussed — leadership challenge sits.
Executive coaching and leadership development have, for most of the past two decades, operated as periodic personal interventions: reflective conversations, 360-degree feedback, off-site workshops, individual development plans. These approaches retain genuine value. They create space for self-awareness, perspective and personal insight that remains important.
But they were designed for a different operating environment. They were built around the assumption that leadership effectiveness was primarily a function of individual capability and behaviour — that developing the leader, in isolation, would translate into better organisational performance. That assumption is increasingly strained.
MIT Sloan’s 2026 AI and Data Leadership research highlights a structural dimension of this challenge: organisations are creating new roles — Chief AI Officers, AI operations managers, human-AI interaction specialists — without establishing clear reporting lines, decision rights or governance structures around them. The result is that AI accountability is diffused rather than held. McKinsey’s 2026 State of Organizations data found that one in six organisations has no clear senior owner of AI at all.
1 in 6 organisations have no clear senior owner of AI responsibility. Where ownership is absent, accountability defaults to no one. (McKinsey, State of Organizations 2026)
Bain’s research on AI transformation failures adds a further layer. AI-related reorganisations are underperforming other types of organisational change — not because employees fail to understand what is expected, but because leaders are not helping them work differently. Fewer than 40% of employees in AI transformations feel the scope and rationale of change were clearly communicated. The failure is not technological. It is human and leadership-level.
The implication is significant. Organisations can no longer develop individual leaders in isolation and expect collective performance to follow. At senior levels, the focus must shift toward helping leadership teams operate effectively as a system — one that can make, challenge, escalate and own decisions in environments where AI is simultaneously shaping the pace and the content of those decisions.
Concerning AI Operational Practices
Consider a CEO running a professional services firm of 150 people. Her finance director is using AI to model cashflow scenarios she hasn’t reviewed. Her sales team is acting on AI-generated pipeline recommendations that feed directly into resource allocation. Her operations manager has automated a client onboarding process that previously involved a human quality check. None of these decisions were made by the leadership team. Each was made by the technology — or by individuals acting on its outputs without escalation.
This is not a hypothetical. It is the operational reality in a growing number of mid-market businesses right now. The question is not whether AI will change how decisions are made in these organisations. It already has. The question is whether leadership governance is keeping pace.
WHAT THIS REQUIRES IN PRACTICE
From periodic coaching to dynamic decision architecture
McKinsey’s State of Organizations 2026 is explicit about the leadership response required: the distinctly human dimensions of leadership — judgement, empathy, purpose, accountability — become more important as AI handles more of the execution, not less. Deloitte describes this as building a “human edge” — expanding from cybersecurity to disinformation security, from process governance to digital trust.
In practical terms, helping leadership teams build this capability means working on five interconnected challenges:
- The speed and quality of decisions under conditions of AI-generated options and compressed timelines
- Governance structures around AI-informed choices — who approves what, and on what basis
- Team alignment in ambiguity, where AI is producing different recommendations to different functions
- Clarity of accountability, ensuring that human oversight is explicit and not dissolved into the system
- Confidence without overreliance — building leaders who use AI as a trusted adviser, not a substitute for judgement
This requires moving beyond coaching as a periodic personal intervention toward a more dynamic advisory role within a decision architecture. It means working alongside leadership teams in the context of real decisions, real governance structures and real operating conditions — not in the controlled environment of a development workshop.
Bain frames the leadership signal clearly: “CEOs who actively use AI, question its outputs, protect disciplined experimentation, and maintain human accountability set the example for how their organisations adopt the technology. Leaders cannot delegate this. The organisation will take its cues from what they do, not what they say.”
THE OPPORTUNITY
Competitive advantage will not come from access to AI — it will come from the quality of leadership exercised with it
The organisations that succeed in the next decade will not necessarily be those with the most advanced AI tools. They will be the ones whose leaders can exercise sound judgement, maintain trust and make accountable decisions in environments of increasing speed, complexity and ambiguity. McKinsey is direct: the shift from short-term resilience to “sustained productivity and long-term impact” will be led by those who treat people, behaviour and culture as the engine — with technology as an enabler, not a replacement.
For mid-market leadership teams, this represents both a risk and a significant opportunity. The risk is that governance gaps and accountability voids — already present in the majority of organisations — compound over time as AI becomes more deeply embedded in operational decisions. The opportunity is that leadership quality, when genuinely developed in this new context, becomes a durable source of competitive advantage in a market where most competitors are still treating AI as a technology project rather than a leadership challenge.
The question for every leadership team is no longer whether AI will change how decisions are made. It is whether their approach to leadership is evolving fast enough to keep pace.
ABOUT MERIDIAN VERITAS
Meridian Veritas works with executive teams in mid-market businesses to build the leadership capability, governance structures and decision frameworks needed to lead effectively in AI-enabled environments. We do not offer generic frameworks. We work inside leadership teams, on the decisions that matter, in the operating conditions they actually face.
If you want to understand how ready your leadership team is for the Age of AI, contact Jeremy Cohen at jeremy.cohen@meridianveritas.com.