You Will Never Have Two Free Days

For the better part of twenty years, I've had the privilege of helping leaders and teams from hundreds of organizations build their capacity to innovate. On campus, in a couple dozen countries, with teams of every size and stripe. It is the work I hold dearest, and I still can't quite believe someone lets me do it.

(If you found me through the AI stuff, a quick word. Long before I was "the AI guy," I was the guy on the design school's faculty running the innovation bootcamps, and everything I now believe about working with AI grew out of that soil. The AI implications are about six paragraphs away, so stick with me.)

One organization in particular has invested in building their people's innovation capacity more seriously than almost any I know. Cohort after cohort, for about a decade now, they keep sending their people, and I keep getting to spend a week with a new room full of innovators in training. I just wrapped the latest one. I came home fired up.

We're always tweaking the program at the edges. But there's one module that has always gotten more of our attention than any other, and if I'm honest, we've never been fully happy with it. We call it "take it home." It's the bridge between the retreat and Monday morning, and that bridge is where innovation programs go to die. Doing this work in a room designed for it is one thing. Doing it inside someone's actual job is a very different thing.

For most of those years, take-it-home opened the same way. Before you leave this room, open your calendar and block two full days in the next month. Not two hours. Two days. You are going to run a complete design cycle on a real problem, and you will never do it unless the time is already spoken for.

We were proud of that opening. It came from hard experience. In the early years (before we invented "take it home") students would leave inspired, go back to work, and nothing would happen, because nothing ever happens in the gaps of an executive calendar. So we made them build the gap on the spot. And it worked. People who pre-blocked two days actually did the cycle. We had the follow-up data to prove it.

Standing in front of this cohort last week, I saw something I hadn't seen in front of any cohort before it.

The take-it-home formula worked exactly once.

Professionals dutifully blocked two days. They did the work. And then they never had two free days again, for the rest of their careers. Which meant the lesson they actually took home was the one we never intended to teach: innovation is a thing you do when you have two free days. Since that condition never recurs, neither does the innovation.

We had solved the immediate problem and installed a permanent one.

The most expensive advice I've ever given

I have been on the wrong side of this for a long time, in print.

I've written that you should schedule unscheduled time. I've written that the most productive AI leaders block time to disconnect. I have told thousands of people, from stages and in workshops, to carve out time to experiment. Some of that advice still holds, and I'll come back to which part. But the core of it, the part that says innovation needs its own slot on the calendar, is the pipe dream that keeps a lot of smart people stuck.

The reason I finally saw it is a conversation I had a few days earlier.

We recorded an episode of Beyond the Prompt with a guest I can't name yet. He's the executive coach the Fortune 50 quietly call when the board and the CEO have stopped being able to talk to each other. Not a technologist. Doesn't claim to be. Which is exactly why what he said landed.

He walked us through Robert Kegan's "immunity to change" work. You state the goal (use AI more, innovate more, whatever it is). Then you name what you're doing instead. And the thing everyone says at that second step, the thing he said he says himself, is: I don't have time.

Then he said something I haven't been able to shake. Blocking the calendar is a first-level fix. It treats the symptom. For some people it works. For most it doesn't, because Kegan's point is that a behavior change never touches the big assumption underneath it, and the assumption is what's actually running the show.

So I did the exercise on myself. Goal: experiment more. What I do instead: tell myself I'll get to it when I have time. And the big assumption holding that whole pattern in place, the one I'd never said out loud, is that experimenting is a separate activity from working.

As long as that belief holds, you will always be waiting for a window that never opens.

Six hundred and eighty-seven steps

I had actually proven his point the week before without realizing it.

Claude recently shipped a feature where you can record your screen while you do a task and then ask it to turn what it watched into a reusable skill. I'm in the middle of a multi-week engagement with the senior leadership team of an entertainment company, and the minute I saw the feature I thought: they need this. Then I thought: I have no idea how it works, and I can't teach something I've never done.

So it sat there for weeks — a little splinter in the back of my mind, filed under things to try when I get a minute.

I never got a minute.

Then one Monday night I was doing what I do every Monday night, which is publish my newsletter. Write it, post it to my site, copy it into Medium, fix the formatting, copy it into Substack, fix the formatting again. It's tedious, I've done it hundreds of times, and it had never once occurred to me to do anything about it.

Halfway through, the splinter surfaced. Wait. Could I just hit record?

I didn't know how to set it up. I googled it. Turns out it's a plus button. I hit record and did what I was already going to do anyway. When I finished, Claude told me it had watched me take 687 steps to "syndicate" my post across four platforms (I didn't even have a word for what I was doing; it did), and that it had turned those steps into a skill.

The thing I "never had time for" took zero additional minutes. I didn't carve out anything. I changed how I showed up to work I was already doing — which, I realized days later in front of that cohort, is the very thing my own take-it-home module had been designing out of people for a decade.

What versus how

So in the room last week, I swapped the question.

The old take-it-home was about what: what new project will you launch, what workstream will you commission, what two days will you block. Every one of those asks the participant to add something to a life that has no room for additions.

The new take-it-home is about how: in the work you are already going to do tomorrow, the meeting you're already going to sit in, the report you're already going to write, what posture are you bringing? Are you going to run it the way you ran it last week, or are you going to run it as an experiment?

I don't have the follow-up data yet. I'll share it when I do. But I've watched enough rooms to know when a module lands differently, and this one did. Nobody was calculating whether they could afford two days. Everybody was thinking about Tuesday.

This matches what I keep seeing in the research behind The Human Advantage. The people who get the least from AI treat it as a separate activity, a tool to pick up when there's time. The people who get the most treat it as a teammate they bring into everything. Same technology. The difference is where it lives in their day.

The part of my old advice that survives

My podcast guest caught the tension in this before I did, on air.

If experimenting on live work (and not carving out separate time to experiment) is the answer, what about spaciousness? What about the walk with no phone, the unscheduled Thursday, the nap? Haven't I argued for those too?

I have, and I still do. But they are a different thing, and conflating them is how I got this wrong for so long. Space for nothing restores you. Time for experiments is a trap, because an experiment that needs its own time slot will always lose to the meeting that already has one. Protect the empty space. Stop trying to schedule the experiments. Run them inside the work.

This week

Identify a task you already plan to do at least twice before Friday. Something recurring and slightly annoying.

The first time you do it, hit record in Claude and let it watch. This should require approximately seven additional seconds of your time, and no more. Then do the task exactly the way you always do.

The second time (the other one that was already on your calendar), check whether Claude turned what it watched the first time around into a skill, and try running the skill instead of doing the task yourself.

Two passes you were going to make anyway. One to record, one to test. Zero minutes added. That's a complete design cycle, and it fits inside a Tuesday.

You are never going to have two free days.

You have tomorrow, and you were already going to be there.

How are you showing up in the work you're already doing?

Related: The Bug Is the Brief

Related: Schedule Unscheduled Time

Related: Hit Reset

Related: Don't Use AI. Work With It.

Join over 55,147 creators and leaders who read Methods of the Masters each week.