AI Upskilling Index

Where is the AI skills gap?

For every AI skill: how many open jobs ask for it – and how many courses teach it. Few courses per job = a wide gap. The figures come from German Federal Employment Agency data; the date of the edition you are seeing is shown just below.

Two official sources, one ratio.
7,713job hits, summed over all AI terms
≥50,729course hits, summed over the same terms
6.6×courses per job (median)

Totals across all terms, not a count of distinct jobs or courses: an ad or course that names several terms counts once for each.

Supply per demand – by skill

Courses per open job, by AI skill. The widest gaps are at the top: high demand, comparatively little training on offer.

To the AI Skills Report (demand)
wide gap (< 3 courses/job)balanced (3–8)plenty on offer (> 8)
  1. 1Data Science1,858 jobs · 2,602 courses
    1.4×
  2. 2Computer Vision191 jobs · 277 courses
    1.5×
  3. 3PyTorch126 jobs · 202 courses
    1.6×
  4. 4MLOps104 jobs · 368 courses
    3.5×
  5. 5TensorFlow90 jobs · 356 courses
    4×
  6. 6Process AutomationSearch term: Prozessautomatisierung450 jobs · 1,926 courses
    4.3×
  7. 7Copilot390 jobs · 1,979 courses
    5.1×
  8. 8Artificial IntelligenceSearch term: Künstliche Intelligenz1,635 jobs · ≥10,000 courses
    ≥6.1×
  9. 9Data AnalysisSearch term: Datenanalyse1,204 jobs · 8,598 courses
    7.1×
  10. 10Machine Learning653 jobs · 5,896 courses
    9×
  11. 11Deep Learning236 jobs · 2,306 courses
    9.8×
  12. 12ChatGPT447 jobs · 4,845 courses
    10.8×
  13. 13Large Language Model45 jobs · 534 courses
    11.9×
  14. 14Natural Language Processing22 jobs · 466 courses
    21.2×
  15. 15Generative AISearch term: Generative KI146 jobs · 5,231 courses
    35.8×
  16. 16Prompt Engineering116 jobs · 5,143 courses
    44.3×

Ratio = courses ÷ open jobs. “≥” = course supply exceeds the source's 10,000 count limit.

Sources & method

Last checked 29 September 2026.

Two independent keyword searches of the German Federal Employment Agency (Jobsuche for demand, NOW/Weiterbildungssuche for supply). Hit counts are snapshots and not directly comparable – the ratio is an indicator, not an exact quota. Course counts cap at 10,000 hits. i6eal is not an official body.

Frequently asked questions

What does the AI Upskilling Index show?

For each AI skill, the ratio of training supply to job demand in Germany – i.e. how many courses there are per open job. Low values suggest a wide qualification gap.

Where does the data come from?

From two services of the German Federal Employment Agency: the Jobsuche API (open jobs per skill = demand) and NOW/Weiterbildungssuche (available courses per skill = supply). The date of the edition you are seeing is shown at the top of the page and under “Sources & method”.

Does “wide gap” mean there are too few courses?

Not necessarily. A low ratio means course supply is scarce relative to demand – a signal to look closer. In absolute terms, most AI skills still have more courses than open jobs.

Why is “Artificial Intelligence” marked with “≥”?

The training search counts at most 10,000 hits. Very broad terms reach this cap; the real supply is then larger and the ratio a lower bound.

AI skills that stick.

From needs analysis to training: we build exactly the AI capabilities your business needs, together with your team.