AI Is Quietly Eating Your Week: Kristian Luoma on the Coordination Tax Index 2026 — The Coordination Tax
Leadership Insights

AI Is Quietly Eating Your Week: Kristian Luoma on the Coordination Tax Index 2026

Episode notes

The conversation, captured

Date
August 7, 2026
Participants
Kristian Luoma

The Coordination Tax is a podcast by In Parallel about what actually breaks down when teams stop sharing the same reality. In this special solo episode, host Kristian Luoma — co-founder of In Parallel and co-author, with Markku Mäkeläinen, of the Coordination Tax Index 2026 — walks through the first quantitative measure of how much of a manager's week goes to managing the work rather than doing it. The Index was fielded in June 2026 across 247 managers in the United States, the United Kingdom, Canada, Australia and Germany, spanning nine industries. It reads 31 out of 100 — nearly a third of management capacity consumed by coordination overhead before any real work happens.

Summary

  • The headline number: 16.5 hours a week per manager lost to coordination overhead. That is 41% of a 40-hour week, roughly two full working days, about 760 hours a year, or around $64,600 per manager at a fully loaded cost of $85 an hour.
  • The perception gap is the finding that matters most. Managers self-report losing 21% of the week. Asked to itemise the same hours, they report 41%. The tax is invisible to the people paying it — which is why it never appears as a line item anywhere.
  • Seniority makes it worse and hides it better. Department heads and directors itemise 20.1 hours a week — half their working time — while self-reporting a loss of only 23%.
  • Where the hours actually go: 8.1 hours of status meetings (49% of the tax), 4.6 hours re-explaining context to people who already had it (28%), and 3.8 hours hunting for decisions that already exist somewhere (23%).
  • Tool count is the strongest structural multiplier in the data (r = 0.42). Teams on 1–3 tools carry roughly 13 hours a week; past ten tools it runs to about 32 hours — a 2.4× jump — and 9% of decisions end up living nowhere but someone's memory.
  • The AI paradox: 86% of managers already use AI at work, and the daily users carry more than twice the coordination load of non-users — 20.3 hours a week against 9.1. 91% load context manually before the assistant is useful. The bottleneck moved from model capability to context access.
  • The shared-context dividend is already priced by the market. Managers estimate that an AI with access to their team's shared context would return 4.4 hours a week — 27% of the measured tax, roughly $17,000 per manager per year, before any process change.
  • What they would do with the time back is not rest: 53% would spend it on strategic thinking, 47% on developing people, and only 6% would leave earlier.
  • Three things the data rules out. You cannot train your way out, you cannot buy your way out with another system of record, and you cannot prompt your way out. A structural cause needs a structural fix.

What was discussed

Intro — 31 out of 100

Key insight
  • The Coordination Tax Index is In Parallel's annual benchmark of coordination overhead, scored 0–100, blending each respondent's measured hours with their self-assessed loss.
  • The 2026 reading is 31, and 48% of managers sit in the High or Severe band.

Why we built the Index

Key insight
  • Most research on meetings measures how people feel about them. This measures where the hours actually go.
  • Fielded in June 2026 via Pollfish — 247 managers, five countries, nine industries, 94% of them managing people directly.

Where the 16.5 hours actually go

Key insight
  • Status meetings account for 8.1 hours a week — just under half the total tax.
  • Re-explaining context adds 4.6 hours; searching for decisions that already exist adds another 3.8.
  • The common thread is that none of it is the work. It is the work about the work.

The perception gap — and the dark matter of the P&L

Highlight
  • Self-reported loss is 21%. Itemised loss from the same people is 41%.
  • Coordination overhead never shows up as a cost line, so it is never managed as one — it is absorbed into salaries and called being busy.

Why it's structural, not behavioural

Key insight
  • The variation tracks structure — tool count, work mode, seniority — not individual discipline or meeting hygiene.
  • Hybrid teams run 8 index points above in-office teams, and directors pay 35% more hours than team leads.

Tool count — the strongest multiplier

Key insight
  • Tool count correlates with coordination hours at r = 0.42, the strongest structural relationship in the data.
  • Each additional tool adds another place for context to land and another place it has to be retrieved from.

The AI paradox — heaviest users pay the most

Highlight
  • The people using AI most intensively are carrying more coordination load, not less — 20.3 hours a week against 9.1 for non-users.
  • 91% brief their assistant manually first, and 55% spend 3–10 minutes or more per session re-explaining what the team already knows.
  • A stateless assistant is a contractor who starts over every morning. Faster output per person does not reduce the cost of getting everyone onto the same page.

The shared-context dividend

Key insight
  • Managers' own estimate of what context-aware AI returns is 4.4 hours a week, about 27% of the tax.
  • Roughly $17,000 per manager per year, before changing a single process.

What managers would do with the time back

Highlight
  • 53% would put it into strategic thinking, 47% into developing people.
  • Only 6% would leave earlier — the demand is for better work, not less of it.

Three things the data rules out

Key insight
  • You cannot train your way out — the load is not caused by bad meeting habits.
  • You cannot buy your way out with another system of record — that is the tool-count finding pointing the other way.
  • You cannot prompt your way out — better prompts do not fix missing context.

What to do on Monday

Highlight
  • Start by measuring rather than estimating. The gap between 21% and 41% only closes when the hours are itemised.
  • The structural fix is to make context persistent and available — captured where work happens, shared across the team, and reachable by the AI tools people already use.

Links

The Coordination Tax Index 2026: download the report

Kristian Luoma on LinkedIn: linkedin.com/in/kluoma

Listen

YouTube · Apple Podcasts

Back to all episodes Book a demo

More episodes