AXHUB5 cards
Why do only the
higher-ups see gains?
Executives say "we save 8 hours a week,"
while most employees answer "0-2 hours."
Same company, same tools.
Why so different — overlay the surveys and a picture appears.
Surveys compiled by Platformer (Section · METR · PwC · Workday) — checked against originals
Productivity paradox1 / 5
01
Start with the size of the gap
A survey of 5,000 white-collar workers (Section).
Two-thirds of employees "save 0-2 hours a week."
Over 40% of executives "save more than 8 hours a week."
The chillier number is this one.
40% of employees said they'd "be fine never using AI again."
Source Section survey of 5,000 white-collar workers (via Platformer, checked against original)
Productivity paradox2 / 5
02
Perception isn't evidence
Who's mistaken?
METR had skilled developers use AI tools.
Measured: 19% slower.
Perceived: "20% faster."
Even the direction was reversed.
The executive's 8 hours and the employee's 0 hours
can both be feelings until you measure again.
Source METR experiment (2025-07) — skilled developers, perception vs measurement reversed
Productivity paradox3 / 5
03
The savings flow downhill
Another explanation is "workslop."
A rough draft the top makes quickly with AI
costs new time below, where people fact-check and fix it.
Workday's survey pointed at the same grain.
Saved time was being offset by review time.
The top's savings become the bottom's overtime.
Source Workday survey · the workslop concept (compiled by Platformer)
Productivity paradox4 / 5
04
CEOs' confessions
Widen it to the company level and the picture is similar.
PwC asked 4,454 CEOs across 95 countries.
"It had an effect on revenue or cost": 12%.
"Haven't gained anything yet": 56%.
Same direction as the survey that "95% of pilots leave no mark on the bottom line."
Source PwC CEO Survey — 4,454 across 95 countries (via Platformer, checked against original)
Productivity paradox5 / 5
05
The fix: measure instead of feel
The paradox isn't because AI is incapable.
It's because we trusted without measuring.
What to start on your team is one metric line.
Something like "share of outputs that shipped with no human edits."
One more — ask where the saved time went.
Whether it went to better work or evaporated is what splits real results.
Source The fix direction in Platformer's write-up · connects to No.12 (agent evaluation criteria)
AXHUBclosing
The paradox comes not from tools
but from a measurement gap
Whether executives or employees are right, you'll know once you measure.
Only the organizations that started measuring
can look at the same number, top and bottom, and set the next step.
Source: Platformer — The AI productivity paradox (citing Section · METR · PwC · Workday, checked against originals 2026-07-03)
AXHub card No.10 — all figures follow the original surveys Platformer cited and are logged in the case library (pattern 13).