Skip to content

Blog

GenAI creates more work before it creates less

A German study of IT leaders finds 73% still have the trough of the J-curve ahead of them. Its authors call the next 18 months decisive.

Written by

Measuring AI impact

Published: 11 December 2025 Updated: 2 min read

73% of German IT leaders still have to go through the trough of disillusionment.

The ones who hit the bottom think they failed. They just didn’t plan for the J-curve. Only 27% already get lasting value from GenAI.

A new German study (iteratec et al.) surveyed 223 IT leaders, and the data is hard reading.

GenAI creates more work before it creates less

46% of IT leaders expect their headcount to grow, against 24% who expect it to shrink. You’re hiring more people to make GenAI work, not replacing people with automation.

This is the J-curve. It’s real, and most teams quit in the middle.

Why it gets worse first

You’re restructuring your entire IT organization:

  • Governance and data quality (70% see deficits there)
  • Reskilling IT staff (97% call it necessary)
  • Prompt engineering competency (73% name it a needed skill)
  • The shift of IT to a strategic enabler

Teams in the trough quit. Teams reaching productivity planned for it.

Competency profiles change completely

97% call reskilling IT staff necessary. The core skill is no longer coding:

  • Prompt engineering (73% name it)
  • Critical thinking and structured practices
  • AI system orchestration
  • Data quality discipline

It is an 18-month journey.

What the leaders do

  • Build centers of excellence (44% of all respondents have one, 10% more plan one) coordinating data quality, prompt engineering, security and change management
  • Invest 3.4% of the IT budget on average, while most companies invest up to 2.5%
  • Spend the 18 months the authors call decisive building foundations (70% see deficits in data quality)

The cost of waiting

Today, your competitors see GenAI as a distraction. In 2027, they’ll wish they’d started in 2024.

The teams that plan to be in the trough for 18 months win. The teams expecting instant payoff quit.

Where are you?

  • Peak: pilots spinning, complexity underestimated
  • Trough: “more work than it saves”, the hardest moment
  • Slope: building foundations with a plan
  • Plateau: productivity unlocked

In the trough? You’re on schedule. Most teams quit here. At peak hype? Brace for 18 months of more work before less headcount.

Teams that planned for the J-curve separate from the field in 2026.

What to do

  1. Assess where you are, and be honest about it.
  2. Resource for a 12 to 18 month J-curve.
  3. Build the foundations first: data quality, governance.
  4. Budget like the leaders: 3.4% of IT spend on GenAI.

What separates the teams is planning for the trough instead of expecting to skip it.

Are you resourced for the J-curve? Or are you expecting instant productivity gains and planning to quit when they don’t appear?

Source: GenAI Study 2026+ (iteratec, kobaltblau, Lünendonk & Hossenfelder, VOICE)

Further reading

  1. The highest AI returns come from automation, not time savings Data from more than 2,500 self-reported use cases shows time savings as the most common goal and the most modest return. Risk reduction, agentic automation and broad adoption pay off more. 1 min readShort post
  2. Junior engineers gain the most from AI coding tools A year-long study of 300 engineers at one company found real productivity gains from AI coding assistants, the largest for juniors, and gains that track how intensively teams adopt. 1 min readShort post
  3. Measuring AI impact on engineering: what to count and what to ignore Code volume, pull requests, usage: AI pushes up every number that is easy to count. A Stanford talk and DX data put the measured gain at around 10 percent, and it swings widely from company to company. What to measure instead, where the returns come from, and why things get worse before they get better. 13 min read

All 16 posts on Measuring AI impact

ClientsClient work

Test the constraint in your system

The Factory Readiness Assessment connects this research to your value stream, your data, and the next decision.

Training on this topic: Agentic Engineering for Engineering and Product Leaders
See the assessment

Want to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check