The Work Was Never the Value

I was on a call last week with someone leading the AI transformation at a century-old media company. Sharp, further along than most. She walked me through the list of "AI projects" her operations lead had handed up. One of them stopped me.

When the team books a client reporting meeting, they spend four or five hours prepping for it. The ask on the list was: can we get AI to automatically schedule those five hours of prep onto everyone's calendars?

I said, why would we schedule the work? Why not have it do the work?

She'd already run that line past her ops lead. The answer she got back: "We're not ready for that. That's too much value."

That one sentence is the whole problem with how this industry is approaching AI, said out loud.

Most media companies aren't behind on AI because they won't use it. They're behind because they're pointing it at the wrong half.

Pointing it at the wrong half

Walk into almost any media company right now and you'll find people "using AI." Drafting emails with it. Making a spreadsheet prettier. Speeding up the report. Scheduling the prep.

All of it is real. All of it misses the point.

The lift was never to do the busywork faster. It was to make the busywork stop existing.

Automating the production of a report that should never have taken five hours is not the win. The win is that there are no five hours — and that time goes back to the client, the market, the call. The part of the job that actually moves the number.

Speed applied to the wrong work just buys you more of the wrong work. It feels like progress. It's the same congestion with a bigger engine.

The work feels like the value

Here's the uncomfortable part — the part "too much value" gave away. Teams aim the tool at the wrong half because the busywork feels like the value.

The five hours of prep is visible. It's on the timesheet. In a lot of these shops, it is the billing. So when you offer to make it vanish, the people who do it don't hear relief. They hear "that's my job."

That's not stubbornness. It's rational. You cannot ask the people whose hours are the product to design away their own hours. I learned that the hard way: I once found one of my best people quietly building her own version of an AI tool, and it was deliberately, almost artfully useless — because she was trying to protect her team's jobs. She wasn't wrong to. I'd pointed the question at exactly the wrong person.

You can't ask the people whose hours are the product to delete their own hours.

Deliver first

So stop asking.

The way through isn't a memo, a workshop, or an AI-readiness committee. It's a finished thing in their hands that they didn't have to build or fear.

When I wanted my own strategy team off the production treadmill, I didn't pitch it. I automated the prep and let the packet land before the meeting — more than they'd ever assembled by hand, already done. The reaction wasn't "you took my job." It was "I don't need all that prep time." And a beat later: "could it also pull in our emails? Could it watch the client's site? Could it flag when our contact changes jobs?"

That's the turn. They stop defending the old work and start pulling you toward the new. You can't argue a team into that. You can only hand it to them.

Point it at the right half

So "are you using AI?" is the wrong question. Everyone is. The right one is sharper, and it sorts the winners out:

What are you pointing it at — and what work are you protecting?

Point it at the right half and the whole equation changes. One team serves more clients. The newsroom clears the rote. The hours flow back to judgment and the relationship — the half a machine still can't reach.

Point it at the wrong half and all you've bought is a faster way to stay exactly where you are.

The tool doesn't choose. Leadership does. And the first act of leadership here is the uncomfortable one: deciding that the work everyone has been protecting was never the value to begin with.