Good governance is not simply about making the right decision. It is also about being able to demonstrate that the decision was reached in the right way.
That was one of the strongest themes to emerge from the Risk Coalition's latest Risk Matters roundtable, which explored The Art of Principled Decision-making with Roger Steare, the Corporate Philosopher. The discussion brought together experienced non-executive directors, risk professionals and governance practitioners to consider a deceptively simple question: how do boards make good decisions when the answer is neither obvious nor contained in the rulebook?
At the heart of the discussion was a framework comprising three complementary "voices": Love, Logic and Law. Love asks who may be helped or harmed by a decision. Logic asks what the facts are, what assumptions are being made and what outcomes are sought. Law asks what the rules require, what promises have been made and whether the outcome is fair. It is deliberately simple. But the discussion suggested that simplicity may be part of its value.
Compliance is the floor, not the ceiling
Boards are familiar with the regulatory and legal dimensions of decision-making. Yet an important distinction emerged between asking "What are we permitted or required to do?" and "What should we do?"
As Roger Steare put it, "compliance is the floor, conscience is our ceiling". The supporting research presented to the roundtable suggested that behaviour changes significantly when people enter the workplace: compliance and rules become more dominant while consideration of the human impact of decisions can become weaker.
This matters because regulation cannot anticipate every situation a board will encounter. The recently published Institute of Directors report, Business. A force for good?, discussed during the session makes a related point: regulation cannot substitute for good judgement, ethical leadership or effective governance. It also argues that moral agency ultimately rests with individuals rather than an abstract corporate entity.
For NEDs, the implication is important. Compliance should be a starting point for challenge, not necessarily its conclusion.
From decision-taking to decision-shaping
Another theme was the distinction between shaping a decision and formally taking it. Too often, the board is presented with a carefully framed recommendation after much of the substantive debate has already occurred within management. Alternatives may have disappeared from the paper, assumptions may be implicit, and the board can find itself approving rather than genuinely deciding.
The practical response is not for NEDs to rerun management's work. It is to ask management to show its workings. What alternatives were considered? What assumptions underpin the recommendation? What evidence contradicts it? Which stakeholders will bear the consequences? What would have to be true for the proposed course of action to be wrong? Those questions turn challenge away from second-guessing management and towards testing the quality of the decision-shaping process. They also address one of the discussion's recurring concerns: boards can place legitimate trust in management, but that trust needs to be capable of verification.
Good challenge needs time and permission
Of course, better questions achieve little if the board environment does not allow them to be asked. The roundtable challenged the familiar problem of overcrowded agendas and insufficient time for genuinely difficult decisions. One example involved an executive meeting expected to consider ten decisions in 90 minutes. For a material strategic or ethical judgement, nine minutes is unlikely to permit meaningful contribution, let alone genuine dissent.
This brings the role of the Chair sharply into focus. Does the agenda create time for the decisions that matter most? Are alternative views actively elicited? Can a director challenge the framing of the question, not simply the recommendation? And is constructive dissent regarded as useful rather than obstructive?
MoralDNA chart: The MoralDNA chart above illustrates how the workplace is associated with a stronger preference for Law - following rules - and a weaker voice of Love - considering the impact of decisions on others.
The discussion of the Post Office Horizon scandal was a sobering reminder of what can happen when boards cease to challenge underlying assumptions and evidence. Crucially, participants also recognised that moral concern alone is insufficient. A director may ask the right question but still receive an apparently logical answer founded on incomplete or unreliable information. Good governance therefore requires both empathy and scepticism: trust, but verify.
Record the reasoning, not just the answer
A particularly relevant challenge for boards was the quality of the decision record. The principle is straightforward: significant decisions should leave a trail showing not merely what the board decided, but how and why it reached that conclusion. What information did it rely upon? What alternatives did it consider? What assumptions did it challenge? What stakeholder consequences were discussed?
This does not mean producing a transcript of board debate. Nor does it mean eliminating the possibility of mistakes. Boards operate with uncertainty, and sometimes a carefully reasoned decision will still produce a poor outcome. The more meaningful governance test is whether the board can demonstrate that, based on the information reasonably available at the time, it approached the decision with rigour, appropriate challenge and good intent.
AI should challenge judgement, not replace it
Finally, the discussion turned to artificial intelligence. The emerging proposition was striking: the most valuable role for AI in important board decisions may not be to provide the answer, but to improve the questions. Used as a "digital tap on the shoulder", AI might identify an untested assumption, prompt consideration of an overlooked stakeholder or expose a perspective that has disappeared through groupthink. But participants were clear about the danger of outsourcing moral judgement to systems trained on historical data that may themselves reproduce embedded biases. The distinction is fundamental. AI can support better human judgement; it should not absolve humans of responsibility for exercising it.
The challenge for NEDs
Perhaps the most useful conclusion from the discussion is that principled decision-making does not require boards to become debating societies or moral philosophers. It requires disciplined curiosity. Before approving the next significant decision, NEDs might simply ask: Who could be helped or harmed? What do we know, what are we assuming, and what might we have wrong? What are our legal and wider obligations? What alternatives have genuinely been considered? And, ultimately, why do we believe this is the right thing to do?
Those questions will not guarantee that every board decision is right. But they make it considerably more likely that the board will make its decisions the right way.
Roger Steare is the author of Ethicability. MoralNDA reveals our moral values and the way we prefer to make decisions about what’s right and you can take the survey here.