Matching
From a skill need to a programme you can book.
Every suggestion carries a rationale an HR director can defend in front of the board and the works council.
Nine disclosed criteria, reweightable per initiative
Industrial automation technology
91
4
1
Illustrative sample scenario. No real customers, no real providers.
Product view
A suggestion as the Company Workspace shows it
The view does not just say that a programme fits. It says why it fits and what it leaves out.
Industrial automation technology
91
Match score
- Technical academy
- Hybrid, alongside work
- 6 months
- October
- Stuttgart and remote
Skills covered 4
- Control systems
- Robotics
- Industrial communication
- Fault diagnosis
Left open 1
- Occupational safety
The rationale, in plain words
This programme is proposed because it covers 4 of 5 target skills for the role Automation technician, runs alongside shift work and starts in October. What remains open is Occupational safety.
Further suggestions from the same run
Every target role gets its own list, sortable by score, date, location or cost.
Digital quality inspection and machine vision
87
High-voltage qualification, levels 2 and 3
84
Condition-based maintenance and sensors
82
Production data and shop floor metrics
79
MES and digital control stations
76
Explainability
How the score is built
The weighting is disclosed and visible inside the workspace. A score nobody can recalculate is a score nobody can stand behind.
Coverage of the skill gap
30
Fit with the target role
20
Skill profile
10
Format
10
Timing
10
Duration
5
Location
5
Working and shift model
5
Cost
5
What that means in practice
Skill coverage outweighs every convenience criterion.
The softer criteria only decide once two programmes are level on substance.
Each criterion is reweightable per initiative, for instance on three shifts with no week to spare.
A score nobody can explain does not survive the first question from the works council.
In detail
What is calculated and what a model contributes
Everything decidable is calculated and filtered. A language model is used only where texts have to be compared and labels reconciled.
Deterministic, recalculable
same data, same result- filters on target role, unit and role cluster
- availability, duration and how many weeks people are tied up
- location, distance and whether overnight stays are needed
- cost and whether the format works with the shift model
Model-assisted, always reviewable
every mapping stays visible- normalising labels so that control systems and PLC engineering count as one skill
- mapping between an in-house role profile and a classification standard
- reading provider programme descriptions for the content they actually teach
- writing the rationale as one sentence that reads without prior knowledge
3,039
occupations in ESCO, plus 13,939 skill concepts
1,016
occupations in O*NET, US Department of Labor
Limits
What matching does not do
It does not decide. QualiShift proposes, ranks and explains. The company picks the programme, not the algorithm.
It does not rate people. The score uses the skill profile of a role cluster, never a performance assessment of an individual.
It stays correctable. Every weight and every skill mapping can be changed in the workspace, visibly in the result.
It gives no funding advice. QualiShift does not assess eligibility or entitlements and files no applications.
Next step
One programme is not yet a measure
The suggestion becomes a learning path, comparable needs become a cohort, and that becomes a request with a date against it.