Directors, put your body on the line
AI’s going to break Boards that don’t adapt around its precise strengths and weaknesses: how good and how bad it is, all at once. It’s going to be hardest for high-performing Boards.
2022’s strategy approval meeting. Diligent analysis, option appraisal, stakeholder testing and endless redrafting has produced an artefact. It brings the weight of a group, the flavours of a culture, the contradictions of the still-unknown. And as the Board discuss, each Director feels it differently. They watch their colleagues; what is working, what isn’t meeting their sense of just-right. How do they feel? And then, think: in this collective, what do I know, think, and feel? Agreement is reached, exhaustingly or easily.
Now look forward to the near future. The strategy AI built is just so damned good. It’s well argued, it surfaces the assumptions and the compromises, the choices to be made and the need to review what data at what future moment. It’s rigorous, clear, it’s just perfect. But no-one’s falling in love with it. And no-one will push through and struggle to deliver something they don’t love.
How does a Board approve?
The dominant story is that only one thing matters: rationality, evidence, analysis, expertise, rigour. But the scale of debate around any Board table immediately says this is nonsense. Boards do not reach agreement by identifying the correct answer. That 2022 Board argued about direction, compromised, reversed their position, doubled down. They weaved all of their perspective into one cloth, that they could all wear with pride.
Next year, it will arrive off the peg, ‘one size fits all’. It might be fabulous. It might fit everyone just right. But that’s it; it gets worn because no-one can object to it. The difference shows up a few months later. Something shifts, and implementation doesn’t look rosy. A Board that argued its way to agreement has an embodied memory, a collective context, with which to rise to the unknown. The Board that signs off the AI document has a document.
published a piece in Forbes last week arguing that AI isn’t a technology you roll out, but a stress test on the organisation. It shows leaders where decisions are unclear, where workflows are bloated, where meetings are performative, where managers are overloaded, and where activity has been mistaken for value.
He’s drawing on Dr Richard Claydon, whose interest is the organisational mechanics nobody can see. In , he names a ‘serve’ layer: ‘the work of integration, responsiveness, translation, coordination, interpretation, repair and stakeholder care’. Claydon suggests that’s where the load lands when everyone starts producing faster. Somebody still has to decide whether the output is right, useful, risky or misleading. Ferrazzi’s pertinent question is ‘Where have we mistaken the artifact for the outcome?’
They are writing about management. I’m applying the same logic to boards, who:
are the last place judgement is applied before a decision binds
carry the ultimate accountability – they cannot pass anything up to a higher authority
have thinner contact with the organisation than anyone else involved
Bodies on the line
Recently, I was workshopping strategy with a Board who were stuck. Mission vs money vs culture vs time vs government. The rational answer was obvious; AI would’ve got to it in minutes. But they couldn’t, for good reason. It cut across too much: values, fears, motivations, revenue, stakeholders, community. Pure rationality could not encompass all of those.
I asked them to stand up and line themselves along an axis describing the tension. Across the room, pick your position. Put your body where your view is. Most of them ended up in the left third of the room, but one Director stood alone at the far end.
That let us talk about something we could not have talked about sitting down. What a dissenting voice is worth. What the Board owes someone who argues vehemently from a minority position. What care they owe the dissenter. What they would lose without them.
And that they still have to cohere, and that the coherence has to be built in a way that respects the person standing apart.
Nobody’s mind changed in that half hour. What changed is that every person in the room had been treated with care, and had watched everyone else be treated the same way. When the decision comes it will be easier, and it will hold.
Getting people out of their chairs and across the room is not something an AI strategy tool can do. Neither is noticing that the room has gone quiet, or that the person at the far end is being talked past.
What a Board is doing when it argues
Complexity generates frustration. NFPs balance multiple contexts, an involved theory of change, diverse stakeholders. There’s no straight line. A Board with the required diversity of perspectives is unable to find a perfect answer that satisfies everyone. That unhappiness isn’t a defect in the process but a signal that optimal doesn’t mean perfect.
In the 2022 Board, the route to the optimal solution involved debate, pushback, compromise. The Board cohered around the decision precisely because it was taxing to reach and partly because they’d watched each other struggle to get there.
As the Board weaves, the warp is what’s fixed (purpose, regulation, constitution, finance…) and the weft is what the Directors pass through it. Reluctance, experience, the stakeholder’s challenging question, the thing we tried two years ago. The cloth needs both – and the material that gets pulled off the loom has a specific pattern, texture and strength to it.
But as the AI takes over, the strategy arrives as finished cloth. Nothing loose, no seam to pull, no thread left hanging. The tool’s weaving is invisible, literally unarguable. The AI models will have baked into them a set of assumptions based on their makers. First, it has a specific way of processing knowledge: binary, rational, linear. Second, the training set of strategy is overwhelmingly based on late 20th Century corporate thinking: Porter, McKinsey, Harvard Business Review.
Claydon put it plainly: ‘If you remove all friction from an organisation, you may also remove its capacity to adapt.’ A perfect strategy is frictionless. The argument in the boardroom may be the last friction left in the process, and Boards need to guard against losing it by accident.
The chief executive is alone with it too
None of this is a critique of the AI-enabled Executives or of the AI tool itself. The CEO didn’t cut a corner, they arrived at the Board with an excellent strategy; but they haven’t wrestled with the emotions either. The CFO gave the robot the numbers. The CMO gave it the demographics. Something that pleased them enough came back, and up it went to the Board.
The trust between a Board and a CEO is made of working through things, not passing upwards and signing off. The fix isn’t better governance, but embracing the mess. Asking What don’t you believe? What do you mourn about option B? It isn’t what we’ve historically asked the CEO for. But in future the Board needs to actively invite the CEO’s emotion back into the conversation: asking for courage and granting safety.
So what changes?
A good Board, handed a strategy it can’t fault, will reach for what it’s already good at. What twenty years of AICD training has told them to do. Test it harder, question the assumptions and the evidence. The AI built the strategy to withstand this test, and it will pass. What has to change is the meeting itself. If the strategy arrives finished, the Board must put the argument back in.
Ask the chief executive where they disagreed with the tool. When the tool asked for data, what was hardest to be clear on? What came back that embraced the contradictions between the Exec, and what elided them into a simple answer?
Ferrazzi reports Claydon arguing that AI could become an extraordinary sensemaking tool if leaders knew how to use it that way, helping teams explore multiple perspectives, test assumptions and surface contradictions. I’m suggesting that’s even more true at the very top.
Resist the urge to see a perfect document as an optimal strategy. Resist the seduction of a great argument. Penn and Teller look like they’ve broken the laws of physics but they haven’t, they’ve just made it look plausible.
And finally, create a safe and emotionally literate Board. What is my heart telling me? How would I explain this approach to a beneficiary; never mind the logic, can I say it myself, in my words, to a human it will affect? How can I seek disagreement, tension, discomfort?
The only measure of the garment you’ve woven is to try it on. Can I wear this, out there, where I will be judged? Can they wear it? Can we wear it? The Board of the future will need to take decisions through fuzzy contingent sensemaking with their hearts as well as their heads.
First published in Strategy, Applied.
