Seven priorities, three decks

Priya is CEO of a food manufacturer in Dandenong. A hundred and twenty people, two shifts, a board meeting every second month. The board pack was due Thursday, so on Tuesday afternoon she asked the company's new AI assistant to summarise the priorities for this financial year.

It came back with seven.

Three were from the March offsite deck. Two were from the strategy paper the board signed off in May. Two were from the budget presentation in June. One of them, a second site in Queensland, had been shelved in July. Priya typed back: "Which of these are current?"

The assistant answered honestly. It couldn't tell. All seven were in documents dated this year, and none were marked as dropped.

Then she asked Dean, her ops manager, the same question. He named four. Two of his people were still costing the Queensland fit-out. Nobody had told them to stop.

The assistant wasn't wrong. It was a very fast, very literal reader of what the leadership team had left behind. And so were Dean's two people.

The warning signs

This is where culture on paper usually starts. Not because leaders don't care, but because the leadership room is where priorities get made, and almost nobody has a job that involves carrying them out of it.

Here's what to check for.

1. More than three priorities. Leaders don't lack ideas. They lack fewer priorities. Five is a wish list. Seven is a budget.

2. Each leader tells it differently. Ask your leadership team to describe the plan in one sentence. If the sentences are about different things, the floor will hear different things.

3. Decisions don't leave the room. The minutes say "agreed". Nobody is named to tell team leaders, and there's no date.

4. Dropped priorities are never announced. Things get shelved quietly in a leadership meeting. Everyone below keeps working on them, because silence sounds like "still on".

5. The plan lives in the deck. It's written in slide language: "drive operational excellence", "grow share in key segments". A team leader can't do anything with that on a Tuesday.

6. The plan comes up monthly in leadership and never in team meetings. If the only people who discuss it are the people who wrote it, it hasn't travelled.

One number

This isn't unusual. A 2018 MIT Sloan Management Review study of 124 organisations found only 28% of executives and middle managers responsible for carrying out strategy could list three of their company's priorities.

Those are the people closest to the plan. The further you go from the room, the lower that number gets.

The gap

Here's Priya's business against the five questions.

What the company says: "We have three priorities this year."

What's written down: Seven priorities across three documents, none marked current, none marked dropped.

What really happens: Each manager works on the priorities their boss mentions most. Team leaders work on whatever their manager raised last week.

None of this is a people problem. Every person in that chain is acting sensibly on what reached them. The gap is in what the leadership team didn't write down, didn't drop out loud, and didn't carry.

If you run EOS, Scaling Up or OKRs, you'll already have good rhythms here: quarterly priorities, a weekly meeting, a scorecard. Keep them. The gap we see most often is that those rhythms reach the leadership team and stop there. The rocks or objectives are clear in the room. Whether the team leader on the afternoon shift could name them is a separate question, and it's the one that decides whether anything changes.

The AI test

When AI tools start reading your plans, three things happen.

First, they find every version. Anything on the shared drive counts. An old deck is as real to an AI tool as the current one, unless you tell it otherwise.

Second, they repeat it. Ask an assistant to draft the next all-staff update, a team briefing or a job ad, and it will pull from whatever priorities it found. Priya's assistant would have happily written a recruitment ad mentioning the Queensland site.

Third, as AI agents start acting on their own, planning production, booking work, setting reminders, they act on the plan they can read. If that plan has seven priorities, the agent is chasing seven.

This is the cheapest diagnostic you'll ever run. Ask your assistant for your priorities. Whatever it says is what your documents say. If that's a surprise, your people are probably getting the same surprise, just more slowly.

What good looks like

It isn't complicated, and it isn't a new framework. Three priorities, one sentence each, in words the floor would use. A named person who takes each leadership decision to the team leaders within a week. A stopped list that goes out every time something is dropped, not just at year end. And one place on the shared drive where the current plan lives, clearly labelled, so people and AI tools land on the same page.

The proof isn't the deck. It's a team leader, a week later, saying the plan back without notes, and knowing what their part is.

The connection

Leadership is upstream of everything, so this is where we start mapping where it travels.

• Team leaders get the plan second-hand, in someone else's words. If it doesn't reach them in words they'd use, it never reaches the floor. That's week 4.
• Operations keeps resourcing priorities that stopped. Dean's two people are costing a site that isn't happening.
• Finance still has budget sitting against shelved work, so nobody notices it's stopped.
• Sales keeps promising what was in the last deck it saw. If a rep told a customer about Queensland delivery, that promise is now someone else's problem.
• Customer service catches all of it, which is next week.

The ident check

Yes or no. Run it with your leadership team, out loud if you're game.

1. Could every member of the leadership team write the same three priorities without notes?
2. Could each of them say what their own team's part is in those three?
3. Could each of them name one thing the business stopped doing this year?
4. Was the last dropped priority announced to the whole business, not just the leadership team?
5. Does every leadership decision have a named person who tells team leaders, and a date?
6. Are the three priorities written in words a team leader would use on a Tuesday?
7. If an AI tool read your shared drive, would it find one current version of the plan?

The first three are the recall test. If you get a no on any of them inside the leadership team, there's no point checking further down yet.

Cheap first

1. Sticky notes before anything else. At your next leadership meeting, have everyone write the three priorities on a sticky note before anyone speaks. Costs nothing, takes five minutes. Put them on the table and compare.
2. Write a "stopped" list. Everything dropped, shelved or paused this year, one line each, with a sentence on why. Send it to every team leader. Dropped priorities should be announced.
3. Retire the old decks. Move superseded plans into an archive folder, or add "superseded" to the file name. Then people and AI tools find one plan, not three.
4. Add one line to your minutes. For every decision: who tells team leaders, by when, and in what words.
5. Then the harder fix. Rewrite the three priorities in words a team leader would use. Test them on two team leaders. If they can say them back a week later without notes, you're done. If not, cut the words down again. Keep what works; just make it small enough to carry.

The Monday close

This week, do the sticky-note test at your leadership meeting before anything else on the agenda.

Next week we follow the plan to where it meets the customer first: customer service. Written policies, unwritten exceptions, and what happens when an AI agent has to answer without anyone's experience to fall back on.

Sources

• No One Knows Your Strategy, Not Even Your Top Leaders, MIT Sloan Management Review, February 2018 (the 28% figure)

Questions people ask

Why do strategic plans fail after the leadership offsite?

Usually not because the strategy is wrong. They fail on the way out of the room: too many priorities, each leader telling it differently, decisions with nobody named to pass them on, and dropped priorities nobody announces.

How many priorities should a leadership team have?

Three, one sentence each, in words a team leader would use. Five is a wish list. Seven is a budget.

How do you test whether the plan has reached the business?

Ask the people expected to deliver it to name the three priorities and their part in them, without notes. Start with the leadership team. A sticky note each, before anyone speaks, takes five minutes.