Every innovation director tells me the same thing. We experiment all the time. We fail fast. So I ask one question: What’s your failure fund? How much are you actually allowed to lose this year, on purpose, being wrong? The room goes quiet. Because there isn’t one. Real experiments cost real money and real people, and the opportunity cost is always too high. So everyone agrees experimentation matters, and almost nobody does it. The VP is flying in, finance missed its numbers, and the Post-it that says think about the future of the business never gets read.
The problem isn’t that you’re not thinking. It’s that you have nowhere to think.
Most leadership teams haven’t run out of intelligence. They’ve run out of space. Every minute is accounted for, and the one thing a business runs on — looking up, looking ahead, asking “what if” — has nowhere to happen. We’ve been told to “get out of the comfort zone” so often that the phrase has gone soft. But it still means something: growth lives just past the edge of comfortable, getting there is hard, and almost nobody volunteers for discomfort. So comfort wins by default, every day.
What if discomfort could be made safe?
The sandbox
In software, a sandbox is a walled-off space where you can run code without touching the live system. Break things. Try the stupid idea. Nothing real gets hurt. Mental sandboxing is the same thing for strategy: a protected, facilitated, time-boxed space where a leadership team thinks all the way through three or four things coming down the line at your sector — before they arrive. No budget at risk. No headcount on the line. Nothing to defend at the next board meeting.
It’s the cheapest experiment in the world, because the only thing you spend is attention. You don’t need a million euros to be allowed to fail when the failing happens in your head. You run the experiment that would be insane to fund and find out, at zero cost, whether it’s mad or whether it’s the future. Most are mad. One isn’t. That one pays for the whole exercise.
Why a room, not a report
You could, of course, just commission the analysis. Pay someone clever to go away and come back with the deck. Plenty of firms will sell you that. But here’s what the research on how teams actually think keeps finding: the value isn’t in the finished analysis, it’s in building it together. A team handed a perfect, polished report does not end up with the shared understanding of a team that argued its way through the messy version themselves. The deck is the output. The alignment — the thing that lets a leadership team move fast and adapt when reality changes — is built only by the people who did the thinking, in the act of doing it.
That’s why mental sandboxing happens in a room, with the team, out loud. You can’t outsource the co-construction because the co-construction is the product. The deck you could buy anywhere. The shared reality you can only build.
You can’t prompt an AI on an empty brain
This is where people reach for ChatGPT and assume the thinking is handled. It isn’t. AI is the best sparring partner I’ve ever worked with — I run two of them against each other and let them argue. But the quality of what comes back is governed entirely by what you bring. Arrive empty, and you get fluent, confident, beautifully written slop. The machine doesn’t fill the void; it amplifies whatever you put in, including nothing.
So the sequence matters. Think first, as a team, in a protected space. Then hand the half-formed thing to the machine to stretch and pressure-test. Do it the other way round and you’ve just automated your own shallowness.
There’s a name for the trap. Back in 1983, Lisanne Bainbridge wrote about the “ironies of automation”: hand the routine work to the machine, and the human is left holding the hardest part — the novel call, the judgment — but stripped of the daily practice that builds the skill to make it. Worse, the complacency this breeds is attentional, not motivational. You can’t will your way out of it, and you won’t notice it happening. Forty years on, that is precisely the risk of running strategy through an AI you arrived at empty. The judgment has to be built somewhere. The sandbox is where.
Are you uncomfortable enough?
If I’ve done my job, you don’t leave the room reassured. You leave safely uncomfortable — facing the gap you already suspected was there but were too busy to look at. Most leaders can feel the distance between how fast the world is moving and how far ahead they’re actually looking. A rally driver reads the corner a quarter-mile out, because at that speed there’s no other way to survive. Most boards drive with their nose on the windscreen, looking a year ahead, in a world cornering faster every quarter.
Mental sandboxing isn’t a list of things to do — you already have one of those, and you can’t get to it. It’s the part before the fixing, the part you keep skipping because it’s uncomfortable and there’s never time. Skip it, and the execution never happens anyway. You can’t act on a three-year strategy if you’ve never let yourself imagine the ten-year world it has to survive in.
So: is your thinking stretched far enough? Or is every hour of every leadership meeting already spoken for by today? If you’ve got that covered — genuinely, carry on. But if some part of you reads this and thinks we don’t do enough of that, the gap isn’t intelligence, and it isn’t effort. It’s space. And space is the one thing you can build on purpose.
Mental sandboxing is a concept I’ve been developing with my long-time collaborator Alan Jordan. If it provokes something — agreement, disagreement, or that small uncomfortable feeling — that’s rather the point.
