The 200 Hours You Saved — Where Did They Go?

A while ago, a marketing friend of mine came to me with a confession.

His team had adopted AI this year, he said. Copywriting, image production, data wrangling — everything got faster. At the quarterly review, he proudly put a line of giant text on a slide: "This quarter, we saved 200 hours."

His boss glanced at him and asked one question.

Where did those 200 hours go?

He froze on the spot.

After a long think, he realized the answer was: nowhere. The work was the same work, the people were the same people, campaigns slipped just like before. The saved time vanished like water poured into sand.

This reminded me of something Sandy Carter once said. Carter is the author of "AI First, Human Always," a former executive at AWS and IBM, and now runs her own privacy-first AI company. The cold water she throws on marketers is blunt:

Saving time isn't the goal. Spending the time you saved on something more valuable — that's the goal.

She has an even sharper formulation: saved time is a currency that expires. If you don't spend it, it's gone.

Powerful. I think that metaphor deserves a spot on the wall of every team currently adopting AI.

Half the Companies Solved the Hard Problem and Skipped the Easy Points

Carter once shared a number that I found hard to believe the first time I heard it.

Among companies already running AI in production, she said, roughly half cannot calculate the return on that investment.

Think about that. Deploying AI is the hard part — vendor selection, system integration, process redesign; all the tough bones already gnawed. And doing the math is the easy part: record a number before you start, record another after, then compare.

And the result? The hard part got done; the easy part got skipped.

So every quarterly review turns into a debate contest: some people feel faster, some feel no change, and the arguing gets heated. But nobody can produce a baseline, and nobody can convince anybody.

Carter's advice is refreshingly plain. Before adopting AI, pick one process you run every week and write down today's cost — hours, dollars, in black and white. Then define "better" as an actual number. Then designate one person to sign off and own it.

Can't manage those three things? Carter's exact words: then what you're running isn't a pilot. It's a demo.

AI without a baseline: the busier you get, the less you can explain.

Three Traps, Each More Common Than the Last

Why do so many AI projects technically work, then get axed at budget review?

Carter's diagnosis is three traps.

Trap one: treating hours saved as the outcome. "We saved 200 hours!" — her follow-up question is always the same: what did those 200 hours become? A campaign that shipped? A sales lead? A person moved to a higher-value role? If you can't point to any of those, then those 200 hours created zero value.

Just like my friend.

Trap two: every department inventing its own scoreboard. AI ownership is shifting from technical teams to business leaders — which is a good thing. But Carter has seen too many organizations where the CFO's desk holds six versions of "success," each measured differently, impossible to rank, impossible to aggregate. Someone has to stand up and set a standard the whole company signs off on.

Trap three: too many projects, no owner. Twelve pilots running simultaneously, zero decisions made. Carter puts it bluntly: that's not a portfolio. That's scattering a handful of seeds and waiting for a show.

The Buyer on the Other Side Might Not Be Human

One step deeper, there's a change bearing down on us — deadlier than fuzzy math.

Someday soon, the buyer sitting at the other end of your funnel will be an agent, not a person.

Actually, it's no longer "someday." Right now, there are agents researching vendors for humans, comparing specs, drawing up shortlists — most of the screening is done before a human ever enters the room.

And the way an agent experiences a brand is a completely different story from a human.

It doesn't watch your videos, doesn't smile at your creative, doesn't remember the elaborate booth you built at that trade show. It reads your pricing page, your documentation, your reviews, your structured data — and then decides: does this one make the shortlist the human will see?

Brands used to survive on being felt. Going forward, they'll also have to withstand being read.

I especially love this line from Carter: the brands that win the next cycle will be readable to machines and lovable to humans. If your marketing can't stand up without emotion, you are slowly becoming invisible to the fastest-growing class of buyers.

Measure First, Automate Second

So, back to the opening question: where did those 200 saved hours go?

If you can't answer, don't rush to adopt more AI. First do those three plain little things: record the cost, define the number, name the person who signs off. A baseline can only be captured now — once the process changes, yesterday's data is gone forever.

Measurement first, automation second.

And this principle isn't just about AI. Any "efficiency gain" eventually has to answer the same question: what did the surplus become?

Answer well, and the budget survives.

Fail to answer, and all you saved was time — not value.

Here's hoping your next quarterly review is no longer a debate contest.