I teamed up with the esteemed Steve Bowman of Conscious Governance and Megan Motto of Motto Advisory, to discuss decision making in times of uncertainty. It was a good conversation, and it kept circling back to something I have believed for a long time. The decision is the fundamental unit of strategy. Everything else is packaging.
Here are the main points.
We are wired for a world that no longer exists
The world is more complex and interconnected than it used to be and it is not going to slow down.
Humans evolved to seek certainty and avoid uncertainty. In a slow moving world that was a sensible instinct. If the world your grandparents experienced was more or less the world you experienced, certainty paid a dividend and uncertainty was dangerous. In a world that changes quickly, the same instinct works against you. Chasing certainty is now one of the main reasons boards make slow, late and expensive decisions.
Which is why the certainty merchants do so well
When we face a hard decision we are often tempted to hand it to people who claim expertise or offer certainty. I call them the certainty merchants, and they are very good at what they do, which is selling the feeling of certainty to those who fear uncertainty.
As a result, decisions slow down, they get heavier, and more information stops producing a better answer. Most boards have felt this and put it down to a process problem. It is not a process problem. It is a context problem.
Boards are full of complicated people
Here is an uncomfortable observation. You usually reach the boardroom by being an expert in something. Expertise belongs in domains, where the right answer can be discovered through analysis and the situation favours those with experience or access to specialist knowledge. Boards fill up with people who are genuinely good at solving analytical problems.
Add the personal liability that sits on directors, and the conservatism that comes with it, and you get the pattern I see most often – boards trying to solve complex problems using analytical methods. It does not work, and no amount of additional analysis makes it work.
Explore Context Before Making Decisions
We use Cynefin, developed by Dave Snowden as a decision support framework. What it does is force you to understand the context before you reach for a decision making tool.
Some situations are clear. The chain has come off the bike. There is a known repeatable way to fix it, so most organisations automate these, outsource them, or push them down to the lowest sensible cost. The role of the board is to make sure the safeguards exist, the process exists, and the process is being followed. Less is more here. Be careful though, because the boundary between clear and chaotic is what we call the catastrophic fold. A system with too many rules and too little flexibility does not fail gently. It fails all at once, and it is very hard to recover.
Add more variables and the relationship between them starts to get complicated. The relationship remains known or knowable, so you invest in analysis, find the experts, map cause and effect, lay out the trade offs and then choose. This is where most boards are strongest, and where they feel most comfortable.
Then there is complexity, where the answer does not exist until you start engaging with the system. Ask a question like “how will AI ultimately change my business” and there is no answer sitting out there waiting to be found. The only way to find it is to start doing things. Trying, learning, getting some of it right and some of it wrong. Think of it the way you think about the score of a game that has not been played yet. You can hold a probabilistic view. You cannot know in advance, you have to play the game to find out.
Chaos is a special case, cause and effect are not available to us in any meaningful way so we must take a big build action just to survive. Boards will have crisis management procedures, business continuity plans and in the worst cases, insurance.
And then there is confusion, which is where we cannot agree on which of the above domains we are in. Some people say we have done this before, just do it again. Others say it is a bit different, we need more data. Others say none of us really knows and we should go and find out. Those three sources of energy can create decision paralysis – we don’t act because we are confused about how to act. Either nothing happens, or three parts of the organisation do three different things for the same reason.
The objective changes with the context
This is the part that tends to land with boards.
In the clear domain, the objective should be to reduce cost and risk. In the complicated domain, it is to build value and defend a competitive advantage. In complexity, the objective is to explore and learn. This leads directly to innovation, which means increasing the bandwidth of your exploration and increasing the surface area for luck. When faced with chaos the objective is to survive and innovate.
So the question changes too. Take a familiar one. A vendor wants several million dollars and six months for a large system that will supposedly fix everything. The complicated question is whether we can afford it and whether we will get the outcome. The better question is smaller. What is the smallest experiment we could do to show this is a good idea? Or, what is the set of small things we could do that would mean we never have to buy the giant thing at all?
Decisions in the complex space are really decisions about how much you are willing to spend to learn enough to make a better call, find a new source of value or avoid a risk.
That is adaptation. The alternative is drift.
The difference between genuine adaptation and strategic drift is not the experimentation. It is what surrounds the experimentation. You need a clear sense of direction, including what you are heading away from, and you need a main effort so everyone knows where to focus energy. Once those two things exist, people have the guardrails to explore safely, and the organisation can distribute decision making to the people actually in touch with the situation. The job of leaders is then to support what is starting to create value and move away from what is not.
Without direction and main effort, the same activity is just drift with better vocabulary.
Decision making is a teachable skill and almost nobody teaches it.
Boards train individuals. Humans evolved to make hard decisions in small groups of people they know and trust, the existence of the board is itself the evidence for that. So we train the individual, then expect the group to perform, and the group crashes into untested assumptions, poor communication and thin situational awareness at exactly the wrong moment.
Train the decision making before the high consequence situation arrives, not during it. A few practical things came out of the session:
- Debrief your decisions. Go back and ask what we knew at the time, what we expected to see, and what actually happened. A debrief is a lead indicator of success, not a lag indicator. If you are debriefing, you are learning, and the next decision should be better than the last one.
- Run pre mortems as well as post mortems. Megan made this point well. Before you decide, imagine the thing has gone horribly wrong and ask what caused this failure.
- Put it in the calendar. If the pre mortem and the decision review sit in the annual board calendar, they happen. If they do not, people forget. Then embed the same questions in your board paper templates so there is space on the page for them.
- Revisit old assumptions. Take a complex decision from the past and debrief it properly. Not to find someone to blame, but to see what you assumed, how close you got, what changed, and what that means for the next one.
AI does not change the context
Worth saying plainly, because a lot of boards are working this out right now. From a systems point of view, AI does not change anything. Complicated systems are still complicated and complex systems are still complex. AI does not change the context.
It is a genuinely useful tool in the clear and complicated space, which we sometimes call the ordered domains. AI is far faster than we are at finding patterns in large volumes of material. Lots of precedent, lots of history, more or less the same thing done many times before. The work Apple published under the title “The Illusion of Thinking” shows the shape of it. Accuracy holds up at low and medium complexity, then falls off a cliff. Different tools sit at different levels of scale, but the shape of the curve is the same.
So where there is little precedent, where things are moving quickly, where cause and effect are not knowable in advance, and where there are competing hypotheses and contradictions in the system, AI is not the answer. Organisations replacing human judgement with AI in that space are taking on far more risk than they realise.
We use what we call the AI sandwich. A human at the start, asking why we are using this and what we expect to get. AI in the middle. A human at the end, asking whether the result is what we needed and whether it is accurate. Humans are still contributing most of the intelligence.
From, “The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity” ,Parshin Shojaee and Iman Mirzadeh and Keivan Alizadeh and Maxwell Horton and Samy Bengio and Mehrdad Farajtabar. 2025, https://arxiv.org/abs/2506.06941
Where to start
If you want one place to begin, use your annual strategy review and ask five questions:
- What worked?
- What did not work, and what have we learned from it?
- What did we miss, because we always miss something?
- What do we take out, because it is done or no longer relevant?
- What do we put in, because something new has happened?
Simple, and a great deal better than the usual approach of walking into the next three year planning session as though the last one never existed.
Thanks to BoardPro
Thanks to the team at BoardPro for hosting the session, and for the work they do making the fundamentals of governance free and easy to pick up, particularly for organisations without much resource to spare. https://www.boardpro.com/
Watch the session
The full webinar recording is here. Decision making in times of uncertainty
If any of this sounds like your boardroom, we run the Flow Learning Lab for teams facing high consequence decisions. It is a compressed learning lab covering how to work together, how to decide, red teaming, debriefing, pre-mortems and human sensor networks.
Maybe just get in touch. stevem@aglx.com