I've been trying to work out for a while why good plans so often get quietly overtaken by the quarter they were written for.
Not wrong, exactly. Overtaken.
Because the planning itself is usually fine. Careful, thorough, a lot of good thinking in it.
It's just that most plans I've seen are built for one future, in enormous and loving detail. And then a different one turns up.
I do this myself, by the way, more often than I'd like to admit. One version of next month, planned beautifully. Then next month arrives wearing a different hat.
So the question this programme is built around is a small one, really.
How many futures were you ready for?
Chaos in the Kitchen is the game at the heart of AI-Augmented Scenario Planning, my scenario planning training. It's one of the 7 games.
What the session covers
It starts with the game. Your leaders run a professional kitchen. Bookings are in. Ratings have to hold. Revenue has to keep moving.
And you can see the trouble from here. The weekend rush. That supplier who has form. The one team member who is having a month.
You know all of it is coming. You have no idea what shape it arrives in.
Teams run the kitchen as a business, with ratings on one side, revenue on the other, and a service coming that nobody can see clearly. Staff are short. Supplies are short. Time is very short. Demand keeps moving.
The plans are meant to look good at the start, so everybody may feel quite clever. Then conditions move.
That's the point where the game is designed to stop being about the plan and start being about whether anyone is willing to reopen it.
It's designed so that teams who made reversible calls early keep breathing, and rigid ones have a very interesting second half.
The future in this game is uncertain, but it isn't unknowable. Not if somebody in the room is actually thinking about it.
What the game is built to bring out
The bet behind it: make the future genuinely unclear and people stop describing how they plan. They just plan.
- Who hedges, and who bets the whole night on one outcome and then discovers there was never a plan B, only a strong feeling.
- Who buys options early, and who locks something in round one that may have turned into a liability by round three.
- Why the best-planned kitchen can run out of clock. A gorgeous strategy with no time left to run it, set against the rough plan that moved early.
- Who negotiates while calm and holding something, and who goes looking for a deal at the exact moment everybody can smell that they need one.
- Whether a reversal gets decided out loud or just drifted into, and whether the room listens to the person who says "we should reverse this" or defends the original call because it was theirs.
- Who can tell a bad decision from a bad result. A sound call still loses sometimes. A reckless one still wins sometimes. Watch who judges a colleague purely on the scoreboard, and who puts their hand up for a poor result that came from a good process.
Then the debrief, and this is where AI comes in
The game is the Experience. I'll mark it with one line - "That was the Experience. Let's unpack that." - and we'll go straight into the debrief, which runs on the EPPA loop.
- Patterns. The room names what kept repeating. The plan nobody reopened. The deal made too late. The call judged on the scoreboard.
- Principles. The room connects that to how they plan at work, in their own words. Be ready for more than one future. Judge the decision separately from the result.
- Application. In most training this is the stage that stalls, and it's the only one that changes next quarter. So it's the one AI gets to help with.
There's no AI in the kitchen. The laptops only come out at Application, on the leaders' own real work.
And at Application, it goes like this:
- Bring one real plan. Each leader brings one actual plan for next quarter. Their own, and the more real the better.
- Ask AI for three other futures. AI drafts three futures they hadn't planned for, and the early signal that would tell them each one is arriving. It's the Lotus Blossom idea, more or less: many options on the table before you pick one.
- Mark what can be undone. The leader goes through their own plan and marks which calls are reversible and which aren't.
- Buy one option now, and set a date to decide again. A named behaviour, a real situation, a date. All three, or it didn't happen.
Imagine a leadership team walking into their next planning meeting with four futures on the table instead of one, and a date already agreed for when they'll look again.
What stays human
AI is quite good at writing other futures. It can't tell a bad decision from a bad result.
That one stays with the leader, and honestly, I think it's the most useful thing they'll practise all day.
What your leaders walk out with
- One real plan for next quarter, with three other futures written beside it and the early signal for each.
- Every call in that plan marked as one that can be undone, or one that can't.
- One option bought now, and a date already set to decide again.
- A way to say "this has changed, let's decide again" in a meeting, without it sounding like anybody got it wrong.
A little about the person running it
I've been training founders and leaders since 2019. The founders have been in the US, Canada, Australia, Singapore and India.
Who it's for
- Senior leaders and strategy teams. The people whose calls cost real money and land two quarters later.
- Teams in the middle of something changing, where the job is staying ready for four plans rather than executing one. A volatile market, a fast-growth stretch, a restructure.
- Leadership teams whose post-mortems keep arriving at the same sentence: "we planned for X, Y happened, and we carried on doing X".
- Strategy offsites, on the morning of day one, or just before a real planning session, so the habits are visible and named while there's still time to change them.
When I'll tell you to pick another one
- Operational and frontline teams. This is built for people allocating resources, and it probably won't land the same way with roles built around execution. Another programme would serve them better.
- Groups with nothing real to trade off. The whole engine needs something to spend.
- When the actual problem is execution quality rather than strategic judgment. Different game, sometimes not a game at all, and I'll say which.
The questions people usually ask
"Our leaders already plan well."
That's fair, and they very likely do. This is more about how many futures the plan was built for, which is a slightly different muscle.
"A game is a lot of leadership time."
It is, and I won't pretend otherwise. What I'd say is that the AI step then runs on the plan they were going to have to revisit anyway, so a good part of the time comes back as work they needed to do anyway.
"Can AI really help with strategy?"
With the judgment, no, and the programme doesn't ask it to. With writing three futures you hadn't thought of, and the signal for each, it's genuinely quite useful. The leader still decides what to do about them.
Worst case, we have a conversation and you come out of it with a clearer brief.
Come and see if it fits
Tell me about your leaders and the kind of uncertainty sitting in front of them, and I'll tell you straight whether this is the right programme. I'd rather lose the booking than run the wrong day.
Come and have a look at the calendar
The kitchen won't make anybody better at predicting the future. Nothing really does. It might make them a little more ready for the one that turns up.
With love from Bengaluru, this is Arvindh saying over and out.