Context switching is what happens every time you drop one task to pick up another: the email in the middle of a report, the Slack ping in the middle of a meeting, the AI draft you need to check before you finish your own thought. Your attention doesn't make the jump all at once. Part of it stays behind on the thing you left, and that leftover drag adds up. By mid-afternoon, it is often the reason you feel fried.
Key takeaways
- Every switch leaves part of your attention on the task you just left.
- Right after a switch, you are slower and make more mistakes, even when you saw it coming.
- The damage is in the volume. One switch is nothing. A day of them is exhausting.
- Supervising AI tools adds a new kind of switch to the workday.
- You can't remove all switching, but you can do a lot less of it.
What context switching does to your brain
The leftover drag has a name. Researcher Sophie Leroy called it attention residue: when you leave a task unfinished to start another, part of your mind keeps working on the first one. You end up doing the new task with less than your full attention.
On top of that, the switch itself slows you down. Decades of lab research show that people answer more slowly and make more errors right after changing tasks. Knowing the switch is coming helps a little. It doesn't make the slowdown go away. And the more different the next task is from the last one, the longer it takes your brain to catch up.
How many switches are in a real workday
In a lab, one switch takes a fraction of a second. The problem is how many of them a real day holds.
When researchers at UC Irvine followed office workers for three days, people spent about 11 minutes on one thing before switching or being interrupted. More than half of those stretches got cut short. And when people did come back to an interrupted task, it was usually 25 minutes later, after two other things had happened in between. The workers said that was the hard part: by the time they got back, they had lost the thread.
Modern work turned the volume up. Microsoft's own data shows its most-pinged users get a meeting, email or chat about every two minutes during core hours, close to 275 interruptions a day. If your slump hits at the same time every afternoon, our piece on afternoon brain fog covers the sleep and caffeine side of the same problem.
Why supervising AI adds to it
The workday now has a new kind of switch. You prompt one tool, check another's draft, approve an agent's output, then answer the Slack message about it. Each check pulls you out of whatever you were doing. Someone has to be the router between all those tools, and that someone is you.
Researchers at BCG gave this a name in a March 2026 Harvard Business Review study: "AI brain fry," the mental fatigue that comes from using or overseeing more AI than your head can keep up with. In their survey of almost 1,500 US workers, the people who spent the most time overseeing AI reported the most mental fatigue and the most information overload. Using more tools helped at first, then stopped helping somewhere past three at once.
It is the same afternoon wall that high performers describe on our corporate athletes page: plenty of capacity in the morning, nothing left by 4pm.
Cognitive load vs cognitive overload
Cognitive load is how much a task asks of your attention and working memory right now. Cognitive overload is when it asks for more than you have. Context switching pushes you toward overload from two sides: every switch makes you load a new set of rules, and the residue of the last task keeps taking up space.
You feel overload in two ways. The first is mental fatigue. At some point your brain starts to decide that pushing on isn't worth the effort, and focusing becomes genuinely hard. The second is stress. When people get interrupted, they tend to speed up to compensate. The work still gets done, and done well, but it feels much worse to do. That is why a day of constant switching can be productive and still leave you wiped out.
Seven ways to switch less, including with AI
- Leave yourself a note before you switch. One line on where you stopped and what comes next. When you come back, you read it instead of rebuilding the whole thing in your head.
- Check email at set times. Three times a day is plenty for most jobs. The rest of the time, close it.
- Group similar work together. Put calls next to calls and writing next to writing. The closer two tasks are, the easier the jump between them.
- Protect one block a day. Mute chat, put it in the shared calendar, and treat it like a meeting with yourself.
- Use fewer AI tools at once. Two or three is plenty. Past that, you spend more time switching between tools than getting help from them.
- Review AI output in batches. Let the agent work, then check its results at set times instead of watching every run come in. Turn off the notifications that ask for your review the moment something finishes.
- Finish your thought before you check the AI. Write down your own answer or outline first, then open the draft. You switch once instead of three times, and you can tell faster whether the output is any good.
After your hardest block of the day, take a real break of ten minutes or more, away from the screen.
What people try that doesn't work
| The idea | What actually happens |
|---|---|
| "I'm good at multitasking." | Most people think they're above average at it. The people who multitask the most tend to be worse at controlling their attention. |
| "Multitasking all day trains my brain to switch faster." | Heavy multitaskers are actually slower to switch. Practice doesn't seem to help. |
| "If I work faster, interruptions don't hurt." | You can keep up the pace and the quality. You'll just feel more stressed and frustrated doing it. |
| "Batching my email will fix the stress." | Batching on its own didn't lower stress. Checking less often did. |
| "More AI tools means more output." | Output stops improving past about three tools at once. |
Cut the switches you can, then support your brain through the rest
Some context switching is just part of the job. Nobody gets to close Slack for a week, meetings will still land in the middle of your best work, and someone has to check what the AI wrote. The goal is to cut the ones you can.
For the rest, some ingredients have research behind them for exactly this. L-tyrosine is the raw material your brain uses to make dopamine, a messenger involved in shifting attention from one thing to the next. In a randomized, double-blind trial, people who took tyrosine moved between two competing tasks more easily than people on placebo. That trial used a single dose much larger than the 500 mg of L-tyrosine in a day of NOOMI, which pairs it with seven other ingredients.
For the stress and mental fatigue that come with constant switching, NOOMI is licensed by Health Canada (NPN 80150439) to help temporarily relieve symptoms of stress, such as mental fatigue, and to help support cognitive function, such as mental focus and mental stamina. It also helps increase energy and resistance to stress over time.
The NOOMI formula breakdown covers the eight ingredients and why each is in the formula.
Context switching FAQ
Is AI brain fry the same as burnout?
No. AI brain fry is short-term mental fatigue from too much AI oversight in a day. Burnout is longer and deeper, and it shows up as emotional and physical exhaustion. You can have one without the other. Using AI to take routine, repetitive work off your plate is actually linked with less burnout.
Is a fried brain from context switching the same as brain fog?
They feel alike. People describe AI brain fry as a fog or a buzzing, with slower decisions and headaches. Brain fog has several possible causes, including sleep, caffeine timing and the post-lunch dip, covered in afternoon brain fog: the 7 reasons your brain quits at 2pm.
Can employers reduce AI brain fry on their team?
Yes. In the BCG study published in Harvard Business Review, people whose managers took the time to answer their AI questions reported less mental fatigue, and people left to figure AI out alone reported more. Workplaces that respected work-life balance had the lowest fatigue of all.
References
- Boksem MAS, Meijman TF, Lorist MM. Effects of mental fatigue on attention: an ERP study. Cognitive Brain Research, 2005 (three hours on task).
- Leroy S, Glomb TM. Tasks interrupted: how anticipating time pressure on resumption of an interrupted task causes attention residue and low performance on interrupting tasks and how a "ready-to-resume" plan mitigates the effects. Organization Science, 2018 (the written resume note; tested in 66 and 44 students).
- Masicampo EJ, Baumeister RF. Consider it done! Plan making can eliminate the cognitive effects of unfulfilled goals. Journal of Personality and Social Psychology, 2011.
- Kushlev K, Dunn EW (University of British Columbia). Checking email less frequently reduces stress. Computers in Human Behavior, 2015 (124 adults, three checks a day for one week).
- Albulescu P et al. "Give me a break!" A systematic review and meta-analysis on the efficacy of micro-breaks. PLoS One, 2022 (22 samples, 2,335 people; after draining work, recovery may need more than 10 minutes).
- Sanbonmatsu DM et al. Who multi-tasks and why? PLoS One, 2013 (310 students; 193 of 277 rated themselves above average; heavy multitaskers had weaker executive control).
- Ophir E, Nass C, Wagner AD. Cognitive control in media multitaskers. PNAS, 2009 (heavy multitaskers' switch cost 167 ms larger; correlational).
- Mark G et al. Email duration, batching and self-interruption. CHI 2016 (40 information workers, 12 workdays; batching showed no lower stress).
- Health Canada, Licensed Natural Health Products Database: NOOMI Gummies, NPN 80150439.
Written by the Noomi team.
