FrancePublic sector2015 – ongoingUrssaf National Fund
Machine-learning risk scores for employer contribution audits
Machine-learning targeting supplied more than half of Urssaf's small-business contribution audits in 2023
Problem
Urssaf needed to select employers for contribution-base audits using the information held in its systems. Risk analysis had to identify businesses that warranted closer examination.
Approach
Acoss, now Urssaf National Fund, tested audit plans derived from machine-learning models for SMEs in 2015. It extended the approach to very small businesses in 2016. Regional Urssaf offices subsequently received annual lists of employers with high risk scores.
Results
Takeaway
Machine learning became a substantial source of audit selection. The share of audits selected did not measure how often those audits found an error or fraud.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Finance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Independent study
Sources
Source summaries
The report traced machine-learning audit targeting from SME trials in 2015 to expansion into very small businesses in 2016. It stated that the approach supplied more than half of contribution-base audits for these employers in 2023.
Full details are in the original sources.