AI is making marketing work more intense, not easier. Here are the guardrails to try

Most lean marketing teams adopted AI to get time back. Fewer drafts to write, fewer reports to build, fewer hours lost to the boring stuff.

Some of that is happening. But two recent Harvard Business Review pieces suggest something else is happening alongside it: the work is getting denser. And of all the functions in one large study, marketing reported the highest rate of what the authors call “AI brain fry”.

That’s worth taking seriously, because lean teams have the least slack to absorb it.

I have personally been feeling this over the last few weeks, to the point, last Monday morning I had to tell my SEO team of Grokbots to STOP - cause I couldn’t keep up with their conversation - and real work they were doing.

Is AI actually making marketing work easier?

Not automatically. In AI Doesn’t Reduce Work, It Intensifies It, Aruna Ranganathan and Xingqi Maggie Ye followed about 200 employees at a US tech company for eight months. Their finding: AI didn’t reduce work. People worked faster, took on a wider range of tasks, and stretched work into more hours of the day, often without being asked.

It’s one company, and it’s a tech company, so don’t treat it as a law of nature. But the pattern will be familiar to anyone who has watched a small team “just try things” with AI.

What is AI brain fry, and why does marketing report it most?

In When Using AI Leads to “Brain Fry”, Julie Bedard and colleagues define it as mental fatigue from excessive use or oversight of AI tools beyond your cognitive capacity. Participants described a buzzing feeling or a mental fog, with trouble focusing, slower decisions, and headaches.

They surveyed 1,488 full-time US workers at large companies. Across everyone using AI for work, 14% said they had experienced it. By 26% of those in marketing had, the highest of any function.

The study doesn’t say why marketing sits at the top. My guess is that marketing has a lot of tools, a lot of outputs to check, and a lot of work that’s easy to start. That’s a guess, not a finding.

Two caveats before you build a policy on it. The data is self-reported, and it comes from large companies, not lean teams. Treat it as a strong signal, not a verdict.

What does it cost a team?

The study links brain fry to costs that a lean team feels fast:

• Decision fatigue: people who reported brain fry reported 33% more of it.
• Errors: they reported minor errors 11% more often and major errors 39% more often.
• Intent to leave: 34% showed active intent to leave, against 25% of those without brain fry.

These are correlations from surveys, not proof that AI caused them. Even so, they describe a team making worse calls, more mistakes, and thinking about leaving. That’s a serious price for a few saved hours.

Where does the extra work come from?

Ranganathan and Ye describe three patterns, and each one is easy to recognise.

• Task expansion. Because AI fills gaps in knowledge, people started doing work that used to belong to someone else. Designers wrote code, and others took on work they’d have outsourced or avoided. Colleagues then had to review it.
• Blurred boundaries. Starting a task became so easy that people slipped work into lunch, meetings, and the minutes before leaving. Over time, breaks stopped feeling like breaks.
• More multitasking. People ran several threads at once, with AI handling something in the background. It felt productive. It also meant constant switching and checking.

The Bedard study adds the mechanical part. The most mentally taxing thing wasn’t using AI. It was overseeing it. People with high oversight reported 14% more mental effort, 12% more mental fatigue, and 19% more information overload. And productivity rose when people went from one tool to two, rose by less with a third, then dipped after three.

Sadly, I’ll admit to leaving work, laptop connected to my hotspot, open, sitting on the passenger seat, with an AI still running - just so I don’t interrupt the build.

Is this burnout, or something else?

It’s worth separating them, because the fix differs. The brain fry study found that heavy AI use predicted mental fatigue, but not higher burnout. Burnout is the emotional side: exhaustion, negative feelings about work, feeling less effective.

On that side, there was good news. People who used AI to substantially cut routine, repetitive work reported burnout scores 15% lower than those who didn’t. So AI can ease the draining stuff, as long as it’s replacing toil rather than adding oversight.

Ranganathan and Ye warn that unchecked workload creep can still lead to fatigue, burnout, and weaker decisions over time. The two risks sit side by side.

What guardrails can a lean team try?

These aren’t proven rules. Some come straight from the research, and some are my own reading of it. Try one or two, and see what changes.

1. Cap the tools in any one workflow. The study saw productivity dip after three at once. Pick a limit, say three, and make “add a fourth” a conscious decision.
2. Count oversight as work. If an AI tool needs checking, that’s time and attention. Plan for it the way you would for a junior hire’s work.
3. Add a decision pause. Ranganathan and Ye suggest brief, structured moments before big decisions. One idea from them is to require one counterargument and one explicit link to the goal before sign-off.
4. Batch the checking. They also suggest sequencing: hold non-urgent updates for natural breakpoints and protect focus windows, rather than reacting to every output as it lands.
5. Keep breaks as breaks. Notice the “quick last prompt” before leaving the desk or the prompt during lunch. The research says these habits wear away real recovery.
6. Say where the saved time goes. If you don’t, it gets backfilled. HBR’s advice to leaders is not to rush to fill the gaps left by automation, because it feels punitive and discourages people from finding more.
7. Make it safe to ask. Workers whose managers answered their AI questions reported 15% lower mental fatigue. Those told to figure it out alone reported 5% higher.
8. Measure impact, not usage. Rewarding how much AI people use encourages waste. Pick the outcome, then judge the work against that.

A one-page check for your team

Before you add another tool or agent, ask:

1. How many AI tools is each person overseeing at once?
2. Who checks the output, and when? Is that time planned or hidden?
3. Did our workload quietly grow in the last three months? If so, was it a choice?
4. Where did the time AI saved us actually go?
5. Are people using AI during breaks or after hours, and do we know why?
6. Do we have one decision pause before our big calls?
7. Can anyone on the team ask a basic AI question without feeling behind?

If you can’t answer most of these, that’s the work. The tools can wait.

Next week I’ll cover the other half: how a lean team does more with AI without trying to do everything.