Using AI to Speed Up Old Work? You May Be Quietly Writing a Layoff Memo

Let me start with a scenario.

You run marketing at a company. Early in the year, the boss made the call: we're bringing in AI. Most of a year later, the report card you hand in reads: emails get written faster, dozens of creative variants come out per day, lead scoring has shrunk from three days to half a day.

The boss listens, and nods.

And then?

Then you may not have noticed: hiding inside your own report is a rigorously argued layoff memo. Every "it's faster now" translates to the same thing: the same work no longer needs this many people.

An efficiency-first AI strategy is, at its core, collecting evidence for "hiring fewer people."

That judgment isn't mine. It comes from Liza Adams, founder of GrowthPath Partners. She spent more than twenty years as a marketing executive, and now specializes in helping companies build what she calls a "Human+AI organization." What is a Human+AI organization? Humans and AI each do what they're best at — and together they take on work that used to be out of reach entirely.

Her view is blunt: most teams use AI to make old work faster. That's not transformation; that's strapping a motor onto an old horse cart. Real growth hides in the work you aren't doing at all right now.

A Test So Simple It's Almost Absurd

What counts as "old work made faster," and what counts as "taking on new work"? She has a test — almost embarrassingly simple.

Pull AI out of the workflow and ask yourself: what happens?

If the answer is "a bit slower, but it's still the same work" — the emails still go out, the variants still come out, it just takes two more days — then all you've bought is a faster machine doing exactly what it did before.

If the answer is "we would never do this work at all" — too slow, too expensive, or simply impossible — then that is where you should be placing your big bets.

Think about it. In the past, scoring leads by hand, one person could manage a few hundred a day, tops, and a ten-person team working a week would cover only a fraction of them. So large numbers of leads were never scored at all; they just stacked up in the pool to rot. This was never a job done badly — it was a job you couldn't afford to do. Today AI can sweep your entire lead pool every day, and you're using it to speed up old work. That's like hiring a world-class chef and putting them on leftover-reheating duty.

Speeding up old work is efficiency. Work you couldn't afford becoming work you can — that's growth.

The Real Obstacle Was Never the Technology

So why do most teams stay stuck where they are?

Is the technology not good enough? No. Adams has seen too many companies for that; the problem lies in three moves.

First, buying licenses and calling it a strategy. Tools purchased, budget spent, company-wide kickoff held — and that's it. Did the way people work change? Nobody brings it up again.

Second, treating a belief problem as a training problem. If employees don't believe this thing is useful, signing them up for three courses won't cure it. A problem of belief can't be solved in a classroom.

Third, the stealthiest one: starting from the org chart instead of from the work. First they debate how positions should be set up, how reporting lines should be drawn, how headcount should be decided — and the work itself comes last. That order, she argues, is exactly backwards.

The work changes first, then the roles, and the org chart comes last.

There's another failure mode almost nobody talks about: how experiments die.

Every leader pays lip service to encouraging experimentation. And then? The first experiment fails, nobody reads the report, nobody holds a debrief, and the person who ran it sits in the meeting with egg on their face. The second time around, nobody raises a hand.

Her summary is sharp: when people can share "what didn't work" without being judged, the experimentation culture is alive; when failure is ignored, everyone retreats to the safest old work.

Rebuilding a Workflow: Where Do You Start?

That's the diagnosis. Now for the cure.

What does it mean to "reimagine" a workflow? It means setting the old blueprint aside and asking from zero: if AI had been there from day one, how should this work have been done all along?

Not taking the original five-step process and compressing it into three with AI. Tearing up the blueprint and redrawing it.

Adams has put this entire playbook into practice inside companies, and some of the workflows she built now run at scale in large organizations. Take it apart and you get three pieces:

1. Instructions. The assignment brief you write for the agent: what it should do, what it shouldn't, and the bar it has to clear.

2. Skills. The toolbox you equip it with: pulling data, drafting copy, running analysis — installed one at a time.

3. Human gates. Which steps the machine runs, and which steps require a human to make the call. This is where the real judgment shows: hand everything to the machine and you're flying blind; install too many gates and you're back in the old world.

At the upcoming MAICON 2026, she'll be on site leading marketing leaders to build, from zero, a workflow that "could never have existed before" — one they can take back and put to work with their own teams as soon as it's built. Last year at MAICON 2025, she talked about training a Custom GPT into a teammate; this time, the topic has taken a big step further. Sharing the stage with Adams are more than 50 speakers from the front lines of AI and business.

By the way, don't expect to collect a pile of one-off prompts at the event and carry them home. What she gives you is a complete method of judgment: first identify which processes are worth rebuilding — the yardstick is the test above — then design from scratch. Twenty-five years of experience, compressed into this one session.

The Efficiency Ledger and the People Ledger

Back to that budget meeting at the start.

Same AI rollout, two roads, two ledgers.

One ledger is called efficiency. Every step saves a little time, and when you do the math at year end, this team can be downsized. What AI has helped you prove: with fewer people, the work still gets done.

The other ledger is called reimagining. Deals you couldn't take in the past, you now dare to take; analysis you couldn't afford, you now run every day; moves you wouldn't have dared to imagine have become routine. What AI has helped you prove: these people, not one of them can be spared — and they need to be stronger than ever.

Stop at efficiency, and you're arguing for fewer people. Move toward reimagining, and you're arguing for better ones.

So the next time you report, don't just report "how much faster." Think of the work you've never done, pick one piece, and build it.

And may you be the one keeping score in the new ledger.