Four levers for change.

By Polle de Maagt · First published · Updated · 5 min read

A strategy is finished when people and the organisation work differently. Not when the deck is.

An old railway signal box lever

So you have a plan. A good one. Everybody nodded in the meeting. Now what?

People are slow to change the way they work, and they’re usually right to be. Their current way works, more or less, and nobody has shown them yet that the new way works better for them. That’s why change is the hardest part of any strategy. The plan is the easy part.

“It is difficult to get a man to understand something, when his salary depends upon his not understanding it.”

Upton Sinclair

In 2011 I wrote down four levers to get people moving. I wrote them for social media. They work for anything. Today they mostly get used for AI. But there is one thing to do before you pull any of them.

Learn ittraining, coachingTry itsmall pilotsSpread itcredible usersFrame itrules and examplestell the story before you pull
Learn ittraining, coachingTry itsmall pilotsSpread itcredible usersFrame itrules and examplestell the story before you pull

Before you pull anything: tell it.

The story, and the purpose behind it. Why are we doing this, and why does it matter to the people who have to do it? Not the board’s version. Theirs.

McKinsey’s influence model makes a similar point: people change when they understand the change and believe in it, see others doing it, have the skills, and find the system backing them up. Three of those are the levers below. The first one is the story.

When Ginni Rometty set out to change how nearly half a million people at IBM worked, she started with three beliefs and turned them into nine behaviours. Meetings opened with a story about one of them. Only after that did the behaviours go into promotion criteria. Do it the other way round and the new criteria look like bureaucracy.

Where it breaks: a purpose statement on a poster. If people can’t tell the story in their own words, there is no story yet.

1. Learn it.

Workshops, training and coaching. Not to tell people what the plan is, but to let them practise the new way until it’s less scary than the old one.

Where it breaks: a training day on its own changes very little. People go back to their desk and their old habits are waiting for them.

2. Try it.

Pilots. Small projects with a clear scope and a clear goal, as outposts of the new way. They let the organisation learn, fail where it’s cheap, and decide the next step based on what happened instead of what was promised.

Where it breaks: a pilot that never stops being a pilot. Agree up front what success looks like and what happens when you reach it.

And keep it small. Jan Rotmans calls it a Gideon’s band: five people with complementary strengths, given time, room and a mandate, and not judged on the usual KPIs. Put a pilot under the normal steering group and the system it was meant to change swallows it.

3. Spread it.

Social power. Who inside the organisation can help? Who do colleagues actually listen to?

This is the lever that changed most since 2011, and AI shows why. One global bank mobilised more than 2,000 AI champions: colleagues who visited teams, held office hours and showed one use case at a time. McKinsey’s design rule for them is the interesting part: champions have to be credible users, not influential voices. The colleague who cut a two-day job to two hours and can show you how.

And look at the middle. Harvard Business Review recently argued that AI adoption is made or broken by middle managers, not by the board. They decide whether the new way becomes daily work or a side project. So they’re the first people to win over, not the last.

4. Frame it.

Guidelines and best practices. Guidelines set the boundaries, so people know what they’re allowed to do. Best practices show what good looks like, so people get ideas. You need both: only rules and nobody tries anything; only inspiration and everybody tries something different.

Where it breaks: when the guidelines are written after the first incident instead of before the first try.

The lever underneath.

Sinclair is the reason four levers are sometimes not enough. If people are rewarded for the old way, no training, pilot or champion will move them. Check what you measure and what you pay for before you pull anything else.

LEGO made that concrete. Since 2024, the bonuses of colleagues across factories, stores and offices are tied to one shared emissions-reduction target. Sustainability stopped being the sustainability team’s problem. What gets rewarded gets done. It cuts both ways, though: a single number in a bonus also invites people to hit the number instead of the goal.

And under all four sit leadership, structure and culture. What leaders do, how the organisation is built and what people see as normal decide whether any lever holds. That deserves its own piece.

Try this

Take one change you’re working on now. First ask three people to tell you the story in their own words. Then write down per lever what you’re already doing. The empty one is probably why it’s slow.

Updated by Polle in October 2026. First published in May 2011 as “Change management toolkit: implementing conversations and consumer connect”. Same four levers, without the 2011 jargon, with the story that has to come first, and with what they look like for AI.

Sources: McKinsey influence model; BCG, “Closing the leadership belief gap” (2026); Jan Rotmans on Gideon’s band; LEGO Group; McKinsey, “How to close the agentic adoption gap” (2026); Gleb Tsipursky, “Middle managers will make or break AI adoption”, HBR (2026). Goes well with: Managing expectations and Ripples, not trends.

Your version of this

Tried it, or thinking about it?

Let me know how it went.

You confirm by mail first. Unsubscribe in one click. No tracking.

Check your inbox. We sent you a link. Click it and you are on the list. Nothing there? Look in your spam folder.

That did not work. The email address looked wrong. Try again.

Polle de Maagt

Keep reading

All thoughts →