The Ergon Index · AI · 2026
0
professionals.

Not a survey. A conversation.

0
words exchanged
0
personal answers
0
minutes of reflection
this is what they told us

8
structured questions
1
personalised written report
97%went all the way through
Finding #1 · Top fears about AI in 18 months

The real fear is pace.
Not replacement.

Falling behind on AI
25.1%
Expertise becoming commodity
12.3%
Confident, no fear
5%
Being replaced
4.5%
Colleagues outpacing me
4%
→Falling-behind anxiety is 6× more common than the fear of being replaced.

“I won't have time to skill up on these tools, and I'll fall behind.”

— Senior HR manager
Finding #2 · % who refuse to delegate, by archetype

Each archetype protects
one act of presence.

Not competence. Presence.

0.0%
the final check
the verifier
0.0%
the strategic call
the strategist
0.0%
the sensitive talk
the connector
0.0%
the precise gesture
the craftsman
0.0%
the original idea
the inventor

“That my value becomes less visible because AI does the visible work for me.”

— Executive assistant
Finding #3 · What they actually want AI to do

The volume goes one way.
The judgment stays.

Give it to AIwhere their week goes
Writing (emails, reports)
74.7%
Admin & scheduling
43.6%
Data & reporting
43.2%
Content creation
37.2%
Coordination & meetings
34.8%
Keep for mewhat they refuse to hand over
The final check
54.8%
Strategic decisions
43.8%
Sensitive conversations
39.8%
Precision craft
26.8%
Creative ideation
11.9%
→They hand AI the verbs of volume (write, sort, analyze). They protect the verbs of consequence (verify, decide, listen).
Finding #4 · “If AI gave you 5 hours back, what would you do?”
20.7%
breathe

One in five doesn't want more output.

More of the same26.9%
Something different20.6%
Just breathe20.7%

They want air.

“I'd finally have time to breathe, and to think long-term.”

— Corporate trainer
In their own words

Six voices from the dataset.

“

That my expertise gets diluted, that anyone can do what I do.

— Senior consultant
“

That competitors move faster than me using AI, and I won't even understand how.

— Investor
“

When I diagnose a problem that nobody else has been able to solve.

— Field engineer
“

I'd work on the deep projects I never have time to open.

— Hospital manager
“

That managers turn to AI instead of turning to me.

— Compensation director
“

When I accompany someone on a hard decision and I see the click happen.

— Trainer

The AI industry sells productivity.

These professionals want presence.

That's a positioning gap.

The long read

Four findings, examined slowly.

The mini-film above was the shape. This is the substance — what each number means, who said it, and what the rest of the world's data says next to it.

01
On the pace fear

The replacement narrative is the press's.
The pace narrative is theirs.

Only 4.5% of the 1,185 said “being replaced” was their top fear. The fear that dominated, six times more often, was being left behind.

The press tells the replacement story because it's legible: a job exists, then it doesn't. The pace story is murkier. It doesn't take your job; it makes your skills feel obsolete in real time. ADP's 2026 survey of 39,000 workers across 36 countries found only 22% strongly agree their position is safe from elimination. Researchers have given it a name: FOBO, the fear of becoming obsolete.

When professionals tell us they're scared of falling behind, they usually mean three things at once. They don't have time to learn the tools. The people who do will outcompete them on execution. And by the time they catch up, the goalposts will have moved again.

That fear isn't irrational. Generative AI reached half of U.S. adults in three years — five times faster than the personal computer, twice as fast as the internet. The tools change every few months. The benchmarks change every few weeks.

What's missing in most companies isn't AI access. It's protected learning time.

The lesson

The professionals who'll thrive aren't the ones who use AI fastest. They're the ones whose week protects 20% for learning — and whose managers don't count that as lost output.

02
On the acts they protect

Five distinct acts.
One per archetype.

We didn't hand professionals a list of things AI can't do and ask them to choose. We let them tell us. Five distinct acts kept coming back: the final check (79%), the strategic call (68%), the sensitive talk (60%), the precise gesture (58%), the original idea (46%).

These aren't tasks. They're acts of presence — moments where a named human takes responsibility for the outcome. McKinsey's 2026 research on the human–AI workforce describes the same shift from the other side: as AI absorbs drafting and analysis, “experts apply judgment and ensure quality. In sales, humans handle negotiation. In customer service, humans handle the emotionally sensitive cases.”

What's striking is the symmetry. When something can be automated, it becomes more valuable for a human to not automate it. The final check matters more, not less, when most of the work was done by a machine. The strategic call matters more when the inputs are infinite. The sensitive talk matters more when the alternative is a chatbot.

Each archetype protects a different act because each trades on a different kind of trust. The verifier trades on diligence. The strategist on judgment under uncertainty. The connector on emotional attunement. The craftsman on signature. The inventor on originality.

The lesson

People don't protect tasks. They protect the one act that makes their name worth their title. The work is figuring out which one is yours, then making it visible.

03
On the trade they actually want

The verbs of volume go to AI.
The verbs of consequence stay.

Look at the trade laid out plainly. The verbs of volume — write, sort, analyze, schedule, coordinate — fill their week. The verbs of consequence — verify, decide, listen, sign — anchor their identity.

There's a clean axis underneath. Things that scale want volume: throughput, speed, consistency. Things with a signature want presence: attention, taste, accountability. AI is exceptional at the first and structurally incapable of the second — because the second requires a name behind the output, and AI doesn't have one.

McKinsey estimates AI agents and robots can already automate over 57% of U.S. work hours. That statistic only becomes useful when you can separate which 57%. Our data answers that precisely: it's the volume side. The writing, the admin, the data work, the meeting-running. Not the verifying, not the deciding, not the listening.

The good news is that the professionals already know which is which. They don't need a consultant to map their week. They need permission — and the time — to actually do the rebalancing.

The lesson

They'll hand AI everything that scales — and protect everything with their signature on it. The work that scales has no name behind it. The work that has consequence does.

04
On what they'd do with the time

One in three doesn't want more output.
They want air.

We asked: “if AI gave you 5 hours back tomorrow, what would you do?” 39% would do more of the same. 30% would pivot to something different. 30% just want to breathe.

This last bucket was the most surprising answer in the dataset. The most repeated word in it: breathe. Then: think, walk, family, read.

Microsoft's 2026 Work Trend Index puts the context behind that answer: today's professionals are interrupted every two minutes, up to 275 times a day. They receive 153 Teams messages and 117 emails daily. Meetings after 8pm are up 16% year over year. One in three say the pace of work in the past five years has made it impossible to keep up. They call it the “infinite workday.”

Against that backdrop, “I'd just breathe” isn't a productivity failure. It's a structural alarm. These professionals aren't asking AI for output. They're asking it for air.

Microsoft's own warning in the same report is bracingly honest: without organizational redesign, AI just accelerates a broken system. More output, same overload. The breathing class is the canary in that mine.

The lesson

One in three professionals doesn't want a productivity gain from AI. They want their humanity back. Any AI roadmap that ignores this will accelerate a system that's already broken.

Closing essay

Why this gap exists,
and what to do about it.

The AI industry is built around the word productivity. Tool launches are framed as efficiency gains. Pricing is benchmarked against headcount. Demos open with save 10 hours per week.

But when you actually sit down with 1,185 working professionals — and let them talk, not check a box — productivity isn't the word that comes back. Presence is.

They want to keep the verbs that matter. They want to be the named human behind a decision. They want to breathe. They're not anti-AI — none of them said get rid of it, the majority use it weekly. They're asking for a different deal: AI carries the volume, I keep the consequence. AI takes the writing, I keep the signature. AI handles the throughput, I keep the time to think.

That trade isn't what most companies are selling, and it's not what most workflows are designed for. We still measure people by output, not by what they protected. We still celebrate the saved hour as a freed hour for more work, instead of asking what the work is for.

The companies that win the next decade won't be the ones that automate the most. They'll be the ones that figure out what their people protect, name it explicitly, and resource it. The professionals who win will be the ones who can articulate their act — the one thing AI cannot do in their name — and make it visible to the people they work with.

That's the whole reason this test exists. To help one person at a time answer two questions: which 75% of my work could AI carry, and what's the 25% that's irreplaceably mine?

— Salim Cherkaoui
Founder, Ergon Test

Find yours.

Take the Ergon Test.

10 minutes. Free. No signup until you want the result.

Discover my Ergon

1,185 and counting · ergontest.com