evidoria

← Back to browse

Good practice Imported

Work-Net AI Job & Competency Matching — Korea's National AI Career-Guidance Platform

South Korea · Eumseong · See the South Korea profile

Evidence: Observational / pre–post Top 40% 53/100 · Ask Evidence Copilot about this practice

Korea's labour ministry and KEIS built AI into Work-Net: 'The Work' (2018) matched ~7,600 people to jobs in a year, cutting search time from 10 min to 5 sec. A 2020 upgrade added AI matching across 12,000 competency categories, shared with universities.

7,600 people
People placed into suited jobs in the first year of 'The Work' (Dec 2018–Nov 2019)
5 seconds
Job-search time after login, down from ~10 minutes
12,000
Job-competency categories analysed (2020 upgrade)
2.7 million
Core keywords analysed (2020 upgrade)
Work-Net AI Job & Competency Matching — Korea's National AI Career-Guidance Platform

Details

Maturity
Scaling
Promoter
Korea Employment Information Service (KEIS) / Ministry of Employment and Labor
Period
2018-present
Keywords
career guidance, public employment service, job matching, competency frameworks, AI/big data

Context

South Korea's Ministry of Employment and Labor (MOEL) and its Korea Employment Information Service (KEIS) have progressively layered AI onto Work-Net, the country's national public jobs portal, since 2018.

Objectives

Use AI to generate personalised job, training and benefit recommendations automatically from a user's CV, training history and interests, replacing manual multi-site searching, then shift to competency-based matching using detailed occupational data.

Activities

'The Work' launched in December 2018, logging users in to generate personalised daily recommendations. In July 2020 MOEL/KEIS upgraded the system to AI matching across 12,000 job-competency categories and 2.7 million core keywords drawn from 18 data sources, and released the resulting occupational data dictionary publicly so universities could use it for career guidance and companies could build their own AI employment services.

Results

Per the OECD's Observatory of Public Sector Innovation, in its first year (December 2018–November 2019) 'The Work' helped roughly 7,600 people find jobs suited to their competencies and cut the time users spent hunting across multiple sites from an average of ten minutes to about five seconds after login.

Conclusions

Both sources describe usage and design in detail but do not report longer-term employment-outcome or retention data beyond the first-year placement figures.

Implementation
Indicative cost
Medium (€50k–€500k) — No published budget; an ongoing national government platform upgraded twice (2018, 2020) within existing public-employment-service infrastructure.
Time to results
Medium (1–3 years) — First AI feature launched December 2018; the major competency-matching upgrade followed in July 2020, roughly 18 months later.
Staffing & skills
Ministry of Employment and Labor (MOEL), Korea Employment Information Service (KEIS)

Conditions for success

  • An existing national public-employment platform providing the user base and data (CVs, training history) that power AI matching
  • Public release of the occupational data dictionary, enabling downstream reuse by universities and firms

Common failure modes

  • No published longer-term employment-retention or outcome data beyond first-year placement figures

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful