One
The machine.
You have to be able to use the base tool first.
Most people in an office do not know their own computer. I do not mean they cannot write code. I mean that the ordinary words fail. Ask someone to go to the file system, or to open the settings for an application, and the instruction stops before the task begins, because “file system”, “settings” and “application” are not words they hold.
They save their work and hope it lands in the same place each time, so that they can send it or upload it later. They do not carry a map of where things go. And the arrangement is fragile. If a shortcut is missing, or a toolbar has moved, they are lost. Even a change of browser can do it.
This is not everyone. Development teams and the people who live in spreadsheets are further along. By my estimate, from the rooms I have worked in, it is about six people in ten. It is an estimate and not a measurement.
Where it came from. We did this. For years, IT told people to log a ticket and wait for support. We told them not to try to fix it themselves. I have told users that. We sent them to help chats that did not help, or we took over the screen remotely and fixed it out of their sight. Under all of it was one message: this machine does not belong to you, so do not damage it.
The result looks like a lack of skill. It is a lack of ownership. People are able to learn where their files go. They were told it was not theirs to learn.
Why it matters now. AI tools change the screen constantly. Each new connector, panel, or update is a missing toolbar to someone who has memorized one path through their machine. An enablement plan that skips this is teaching the second thing before the first.
What I do about it. It goes at the very beginning of the first session, before anyone opens the tool. It has two parts, and the order matters.
Permission comes first. I say it plainly:
“this is a piece of plastic and silicon. you cannot break it by typing stuff into it. and we have built in guardrails for accidental clicks. click away. experiment. we can roll back your mistakes.”
Those are promises, so they have to be true. Before I say them to a room, I confirm that the guardrails and the rollback exist. That is one of the things the governance step is for.
Then the map. I teach the file system, at the level of a first course. I explain that applications are programs running on top of an operating system. And I teach a way of seeing: everything on the screen is an object, and every object has a description and a set of actions. A right click shows what the object can do. A left click does the most common thing. I recently taught this to a seventy-five-year-old, outside any engagement, and it held.
A person who memorized a path is stranded when the path changes. A person who holds that model can walk up to a screen they have never seen. If you know a little about how all of this works, you move faster and with less fear.
You have to be willing to fail in this world. That begins with being told that you may.
Two
The time.
Redirect it before it is redirected for you.
When AI gives a person hours back, the question is what happens to the hours.
What I saw. In the engagement behind this site, people who got time back went one of two ways. Some kept climbing. They wanted the next project, the next tool, the next problem. Every one of them already had a list. They had been thinking, for a long time, “if I could just do this, here is what I would do.” The others spent the time on themselves: time online, a rest, a walk.
The difference between the two groups was the list. It was not character.
Whose time it is. I will say this plainly, because it is usually left unsaid. Time at work is the organization’s time. The person has sold it. So freed time will be accounted for, by someone.
The method asks for that accounting openly. The baseline taken in step one is taken again afterward. And, perhaps surprisingly, the result is useful to the person. You know what time you now have, and you can begin to point it somewhere.
The aim is simple. You redirect your time before it is redirected for you, or before it is no longer paid for.
Why people pocket it. Many people at work are overwhelmed. Some feel ill-used, and many feel underpaid. Their lives have often come second to their jobs. To them, an hour released by AI looks like a hundred-dollar bill found on the street. They smile and put it in their pocket.
That is a sensible response to the usual rule, which is that the reward for finishing early is more work at the same pay. Nobody changes that response with a slogan.
What the organization has to supply. An incentive, and an honest account of what AI is for.
On the second point: AI should be presented as something that powers people up, and the organization has to believe it. I do not call this spin. AI is not here to make less work. It is here to make work more useful, and the hours it frees are for the work people never had time to do. An organization states which it believes when it answers the third governance question in step three, and that answer has to be given honestly.
On the first point, the incentives are different in different places.
Where people hold equity, the conversation is easy. Everyone works toward the common goal because everyone is a part owner. Where there is a real mission, people are united by it. And some of the reward is private and cannot be issued by anyone: hard thinking feels good, and the people around you see what you do.
Where there is no ownership and the mission is thin, two incentives remain, and they are blunt. The first is the vote. Early in this change, a person who is proactive about designing their own role after AI gets a say in it. The second is future paychecks.
How a person gets a list. People who already have one need only be asked. People who do not have one can be taught to make one, and this is the part of enablement that almost nobody plans for. It is a second un-training: learning to think bigger inside a role.
The method begins this on day one. The tenth survey question asks for a moonshot. The homework in step seven asks for it again, on paper.
And this can be taught - it has to be taught. Teaching people to dream bigger in their roles is part of the work itself: make the list, and see yourself at six months, a year, five years. The organization writes that timeline for itself in step four of the method. The person deserves the same horizon.
Two lessons carry it. The first: working cross-functionally is ok again. The same don’t-touch era that fenced people off from their computers fenced them into their lanes - stay in your silo was trained as deeply as stay in your inbox, and it gets un-trained the same way. The second is said plainly, because people deserve plainness: you have a better shot at a future if you dream bigger. The list is not homework. It is the odds.
Late in the engagement I built a shared project for exactly this purpose, called MyWork. It does one of two things. It uncovers the list a person has been making without knowing it, or it starts them making one. It asks four questions, each of which seats the person somewhere with authority. Two of them:
“If you were starting this job fresh tomorrow, knowing everything you know now, what would your first 30, 60, and 90 days look like?”
“Step up a level. Imagine the department is yours to steer. What's the one big change you'd set in motion?”
From the answers it drafts a one-page charter in the person’s own words. The charter belongs to the person. I tell them these are their decisions to make.
I should be exact about its standing. I built MyWork in the last weeks of the engagement. I do not know that any member of staff used it before I left, and I doubt it. It is offered here as designed and built, not as proven.
Where the plan goes. In the organization I would design, the person brings their plan to their manager. A manager who is brought a plan owes that person an answer: what can be tried, when, and what comes off their plate to make room. What else is owed, I am still working out. The charter stays private. What reaches the manager is what the person chooses to bring.
Three
The work.
Is it yours?
Before any of the steps, an organization has to answer one question. What counts as work?
Here is my answer to the part of it that AI has made urgent. AI-assisted work is work. The computer did not do it spontaneously. We did it together.
How the discount arrives. It rarely comes as a refusal. It comes as a compliment. I used AI to write the specifications for an engagement, and the response was:
“wow this is really good, i guess you used ai”
Of course I did. I was there to teach these tools, and I use them. That is why I was the one asked to do the job.
Praise that is given and taken back in one sentence teaches a lesson. The person who hears it smiles, and the next time they do not mention the tool. I had the standing to answer. A member of staff hearing the same sentence from a manager usually does not.
What it costs the leader. When AI-assisted work is marked down, people stop showing it. The industry has a name for unapproved tools used in secret - shadow AI. This is the stranger version: the approved tool, used in the shadows, because showing the work costs too much. Use that is hidden cannot be measured. The ledger cannot see it, and neither can the before-and-after assessment of time. Building in public stops. Nobody brings a plan to a manager who has already said the work behind it does not count. The leader who says “it doesn't count” loses the numbers along with the goodwill - and inherits a shadow AI problem no policy will find, because policy was never the cause.
The other way to read it. Using a tool shows that a person is using everything at their disposal to get the job done. That is diligence, and it should be encouraged.
The question to ask. The old question is about how the work was made. Did you use AI? I would retire it. When we take the products of AI-assisted labor into the world, this is what we should all be asking each other:
“do we stand by this - do we own this - is it yours.”
It is a higher standard than the one it replaces. Under it, the tool is no excuse in either direction. Using it does not mark the work down, and using it does not excuse a mistake in the work. And it runs both ways. Staff may ask it of a leader, and each of us asks it of ourselves before anything leaves the building.
Ask whether I used AI when the answer matters - a client, a regulator, a reviewer may need to know how work was made. Just don’t use the answer to decide whether the work counts. Ask what I checked, and whether I stand by it.
What this opens. This is what I mean when I say enablement is about ungating the workplace. The gate being removed is the rule that work only counts when it was done by hand. Past that gate there is collaboration that cost too much a year ago. I make the cross-department introductions before I leave for this reason, and the horizon the method points at is AI joining the work of separate departments. I state that as what the method is built to cause. I do not yet have an example I watched happen.
The personal version of this argument is in “What Counts”, on the parent site.


