Your boss just read a flawless report. Polished, thorough, delivered fast. Problem is, they have no idea whether you wrote it or ChatGPT did. And according to a new survey, neither do most managers. That’s the messy reality of AI at work right now.
What the survey actually found
Monday.com commissioned Nielsen to poll 1,700 directors and senior managers across seven countries. The results landed this week, and they’re awkward.
More than half of senior leaders say it’s now harder to evaluate employees’ work after adopting AI. Not easier. Harder. Nearly half say they spend more time reviewing and guiding output than they did before, because so much of it arrives already polished.
So the productivity promise came true, and it created a side effect nobody planned for: when everything looks great, nothing proves who did what.
If AI at work feels messier than the demos promised, the data agrees with you.
Here’s the breakdown:
| Survey finding | Share of managers |
|---|---|
| Harder to evaluate work after adopting AI | 50%+ |
| Spend more time reviewing output than before | ~50% |
| Changed how they evaluate employees | 78% |
| Excited to improve work with AI | 94% |
| Use AI to inform their own decisions | 46% |
| Use AI to draft performance reviews | 39% |
| Say their org has more tools than before | 57% |
| Say tools became more interdependent | 51% |
Why managers are struggling with AI at work
Michal Miller Levi, head of market research at Monday.com, put it plainly to TechCrunch: AI is moving faster than the organizational processes built to manage it.
The mechanics are simple. An employee does a task with AI help. Output arrives in the manager’s inbox looking finished. Was it two hours of real skill, or ten minutes of good prompting? From the inbox, those look identical.
Meanwhile the managers themselves aren’t anti-AI, far from it. 94% are excited about it. Many know exactly where it helps them most. The friction isn’t enthusiasm, it’s measurement. Their whole job is judging output, and their main instrument for doing that just broke.
The bias problem nobody’s talking about
Now stack this on top of something weirder. 39% of leaders already use AI to draft performance reviews. 46% use AI to inform decisions about their people.
So an AI-assisted manager uses AI to judge AI-assisted work. As TechCrunch noted, that risks funneling more bias into reviews, because language models are prediction machines that repeat patterns from their training data. Research from the Luxembourg Institute of Socio-Economic Research found LLMs rate performance similarly to humans when clear standards are missing, and a 2025 paper found models can show significant bias against certain groups.
Translation: the person deciding your promotion might be judging you through a filter that quietly favors people who write (and prompt) like the average of the internet. It’s worth knowing that before your next review cycle, not after.
How to make your work visible again
Here’s the practical part. You can’t stop your company from using AI. You can change what your manager can see.
- Show the before. Keep the messy first draft, the original brief, the ugly spreadsheet. When you deliver, a one-line “here’s where we started” makes the distance you traveled obvious.
- Narrate your decisions. Two sentences in your update: what you tried first and why you chose this path. AI doesn’t have judgment stories. You do.
- Be the one who checks the AI. Volunteer to review AI output for errors before it ships. Suddenly you’re the person whose judgment everyone trusts, which is exactly what 78% of managers now say they value.
- Document the prompts that worked. When a campaign lands, you want to say “I built this approach” with receipts. That’s becoming a portfolio, not a quirk.
Treat those four habits as basic hygiene for AI at work. The tools changed what managers can see, so showing your thinking is part of the job now.
None of this is resume-padding. It’s making skill legible in a world where raw output stopped being proof of skill.
The skills that now get you noticed
Look back at that 78% who changed how they evaluate people. What they’re weighting instead: how well you write prompts, which tasks you delegate to AI, and how you experiment with new ways of working.
Notice what’s not on the list. Typing speed. Hours logged. Raw production volume, because AI made that cheap.
This lines up with what the hiring market has been signaling all year. Employers say they’ll pay a premium for AI skills, but the premium isn’t for producing more stuff. It’s for knowing what to produce, spotting what the machine got wrong, and owning the outcome. The output is the commodity. The judgment is the job.
There’s also an honest warning here. If your only contribution is pasting tasks into a chatbot, you’re in trouble, because your manager can see that now too. The same feature that hides your effort hides your absence of effort. And there’s a real difference between using AI to write everything versus using it to move faster on work you actually direct.
If you manage people, do this instead
Maybe you’re on the other side of this. Half your inbox is AI-polished work and you can’t tell what’s real anymore. A few adjustments help:
- Review process, not just product. Ask how they approached it. The reasoning reveals skill that the deliverable hides.
- Watch the edges. Whoever caught the AI’s mistake learned more than whoever shipped it clean. That’s the behavior to reward.
- Keep AI out of the judging loop for now. With no clear standards, using AI to draft reviews just averages people against patterns. Until your org has real evaluation criteria, your own eyes are the better instrument.
The survey’s most honest number might be that 57% of orgs are running more tools than before while nobody rethought the process around them. That gap is a management problem, not a technology problem. If you’re a manager getting started with all this, a structured AI adoption plan beats buying another subscription.
AI at work is also changing job interviews
One more angle worth five minutes. If you’re job hunting, assume every interviewer has seen these same statistics. Expect questions about how you use AI, and don’t answer with “I use it for everything.” Walk through one specific task instead: the approach you tried, what the model got wrong, and what you changed. That answer demonstrates the exact judgment 78% of managers say they now screen for.
And if you’re early in your career? This is honestly good news. The “invisible junior” problem, where you grind through grunt work for two years before anyone notices you, gets easier when the grunt work is automated and your judgment is the visible part. Bring receipts to interviews now. It’s becoming as normal as bringing a resume.
What to do this week
Pick your most recent deliverable and spend ten minutes writing down what you decided, what you rejected, and what you fixed by hand. Send two of those sentences to your manager unprompted. Then repeat next week.
The people who thrive in the next two years won’t be the ones who hide their AI use or brag about it. They’ll be the ones whose thinking is visible on top of it.
Source: TechCrunch’s report on the Monday.com/Nielsen survey, published October 21, 2026.
