Visible work

Organizations don’t run on work – they run on visible work.

Dorka Nagy-Józsa, Horizontal Thinking (2027)

We keep asking when AI will take the work. Wrong question, wrong sequence. Horizontal Thinking argues that AI takes something else first - the visibility of effort - and that this, not task removal, is the earthquake, because promotions, trust and salary bands are all priced against effort someone upstairs can see.

01

It takes the visibility first

My company builds assessments, and I've defended them in boardrooms. So in the spring of 2026 I told my own AI agent to complete a set of recruitment assessments in my place. Not prepare for them. Complete them. As me.

It passed them. All of them, consistency traps included - a consistency trap is a pattern, and patterns are its mother tongue. No work disappeared that afternoon. What disappeared was the connection between the output and the effort behind it.

AI doesn’t take the work first. It takes the visibility of effort.

For as long as offices have existed, a polished report meant somebody's Tuesday, and a strategy memo meant a week of somebody's evenings. “Output was a receipt for effort.” You could read the hours right off the deliverable, the way you read them off a hand-knitted sweater.

That receipt is now void.

You've felt it. A colleague sends a flawless six-page analysis and your first thought is no longer “impressive” but “I wonder how long this actually took.” A week, an hour, or a lunch break - you genuinely cannot tell. “The document has gone mute about its own origins.”

02

Organizations run on visible work

Promotions run on visible work. Trust runs on it. Salary bands, reputations, the whole quiet sociology of who gets listened to in meetings - all of it is priced against effort somebody upstairs can look at.

The first symptom is already walking around your office: people hiding their speed. A marketing lead - the book calls her Marta - finishes her monthly campaign analysis on Monday morning, in under an hour, and releases it on Thursday afternoon. Her reason: “If they knew it took an hour, they’d think it’s worth an hour.” She isn't lazy. She is fluent in the old currency, and she has noticed, correctly, that her bosses still pay by the old exchange rate.

We used to perform effort we had genuinely spent, just on the wrong things. Now people perform effort that never existed.

That is not a workflow problem. That is a currency collapse.

And with the visibility gone, something embarrassing steps into the light: a lot of what we called thinking at work was pattern-matching and template-filling. The template just used to take all day, so we called it analysis. Say that to a live audience and the pushback is never “that’s false.” The pushback is “Shhhhh!!”

03

The legibility economy

We built work this way because of a structural bias so old it never got a name: “organizations reward legibility over originality.” A clear slide deck over a messy insight.

Legible means a manager can see it, count it and forward it upward without translation. The actual moment of value is completely illegible. Try putting it in a status update: “This week I had one thought. It was the right one.” I dare you.

 LegibleIllegible
Looks likeA forty-page reportOne thought. The right one.
Shows up asA calendar without white space, being online at 7:40 p.m.Thirty seconds in the shower
EvidenceAttachment, timestampNo attachment. No timestamp.
GetsThe promotion“interesting – can you put it in a deck?”

A bank director I worked with, Hans-Peter, inherited a sixty-page monthly report that took two analysts four days a month. He read page one and the two charts on page fourteen. When I asked why he didn't kill the other fifty-eight pages, he looked at me as if I'd proposed arson: “Then how would the board see how much work my team does?”

“That report stopped being information years ago. It’s a ritual object.” Everyone is playing by the rules: the board grades what it can see. Economists call it Goodhart's law: when a measure becomes a target, it stops measuring. We made visible effort the measure of contribution and received exactly what we ordered - effort theater, performed nightly by honest people who would have preferred to do something useful.

For two hundred years the theater was affordable, because effort and output were genuinely correlated. Then the correlation snapped, and the theater got better. Polish used to be a costly signal; now it's free, so it signals nothing - and the grading system your organization runs on returns the same score for everybody.

04

The homework became ungradable

The economist Michael Spence explained in 1973 why signals work at all: because they're costly. A degree never proved you remembered the coursework - it proved you could endure four structured years. The cost of the signal was the information. Biology knew it first: a clip-on peacock's tail fools exactly zero peahens.

And a cheap signal, per Spence, is not a weaker signal. It’s not a signal at all.

A quick roll call of the dead:

  • The CVDied first. No obituary, no flowers.
  • The take-home taskDied next.
  • The unproctored online assessmentDied on my desk, before the coffee cooled.

Every instrument built on written output produced without observation now measures one thing: access to a twenty-dollar subscription. Universities are buying paper exam booklets again and rediscovering the oral exam. “The scale is what died, not the measurement.”

When clients ask me, weekly now, “then how do we assess anyone?”, I hand back the question they've been avoiding: does your process measure what the person knows, or what they can produce with AI assistance? If it's the second, say so in the job ad and design the process honestly. Watch people work instead of grading their artifacts, sample judgment live, put real consequences back into the exercise.

05

What we actually valued

Whether AI can do your job is the smaller worry. The collapse ends in an audit question: were the parts of your job that make you irreplaceable ever the parts your organization actually valued? Ask it about yourself, not your team.

For a painful number of excellent people, the honest answer is no. Their judgment was tolerated as a charming quirk while their legible output was celebrated and priced into every bonus. Twenty years of that trains a person to spend their judgment budget on making things presentable - exactly what the machine now produces for the same twenty dollars. My own industry is no exception: we print the certificates suitable for framing.

Judgment was always the point.

Some people are about to give a wonderful answer: the quiet ones whose insights never fit the slide template, the analyst who mumbles in meetings and is correct. Their stock is about to reprice violently upward, because when everybody's deck is equally beautiful, the only differentiator left is being right. Others will discover their whole professional identity was scaffolding-shaped. Who is standing over the trapdoor is the next chapter.

06

What to do on Monday

  1. Highlight your last performance review.The one you got or the one you gave. Highlight every sentence that praises legible output: documents, decks, reports, responsiveness, “thoroughness.” What's left unhighlighted is everything your organization officially knows about your judgment. For most people it's under three sentences.
  2. Test one of your own signals.Give an agent your hiring task, your certification quiz or your monthly report template - whichever signal you lean on most. Test a copy, not the live one, and if it belongs to HR, they hear the result first. If it passes, put its replacement on this quarter's plan, before your candidates retire it for you.
  3. Take a marker to your proudest deliverable.Mark every part an agent could have produced. Whatever survives the marker, say it in your next one-on-one. That's the asset. Grow that.
  4. Score your praise.If you lead people, list the last three pieces of work you publicly praised. Clear deck, or messy insight? If the score is three to zero, you're still paying premium prices for scaffolding - and the market price just went to zero.

What survives the marker is the start of a position, not a task pile: position crafting shows how to build one.

07

In the book

This page draws on Chapter 2 (The Scaffolding Falls) of Horizontal Thinking: Leadership after the end of the specialist era by Dorka Nagy-Józsa, PCC, published by Y2Y Ltd. on January 1, 2027. The book has fifteen chapters and two letters from the future, and every chapter ends with something you can actually do on Monday.

Dorka Nagy-Józsa is CEO of Y2Y Ltd., an ICF PCC coach and an international keynote speaker on AI and leadership.

09

Questions

What is visible work?

Visible work is effort that someone upstairs can see, count and forward without translation: the forty-page report, the calendar without white space, being online late in the evening. In Horizontal Thinking, Dorka Nagy-Józsa argues that organizations don't run on work - they run on visible work. Promotions, trust, salary bands and reputations are all priced against it.

What does it mean that AI takes the visibility first?

AI doesn't take the work first. It takes the visibility of effort. For as long as offices have existed, output was a receipt for effort: a polished report meant somebody's Tuesday. Now a flawless document can mean a week, an hour or a lunch break, and nobody can tell which. The book calls this, not task removal, the earthquake.

Why do people hide how fast they work with AI?

Because their bosses still pay by the old exchange rate. In the book, a marketing lead finishes her monthly campaign analysis on Monday morning in under an hour and releases it on Thursday afternoon, because if they knew it took an hour, they'd think it's worth an hour. People used to perform effort they had genuinely spent; now they perform effort that never existed. The book calls that a currency collapse.

Can AI pass recruitment assessments?

In the author's own experiment in the spring of 2026, her AI agent completed recruitment assessments in her place - personality inventory, situational judgment test and all - and passed them. Per Michael Spence's 1973 signaling theory, signals work because they are costly; AI made written, unobserved signals cheap, so they stop being signals. The book's advice: watch people work instead of grading their artifacts, sample judgment live, and put real consequences back into the exercise.

Where does the idea of visible work come from?

From Chapter 2 (The Scaffolding Falls) of Horizontal Thinking: Leadership after the end of the specialist era by Dorka Nagy-Józsa, PCC, published by Y2Y Ltd. on January 1, 2027.

Read the whole argument.

Horizontal Thinking is published on January 1, 2027. The Kindle edition is open for pre-order now; the paperback goes on sale on Amazon the same day.