United KingdomPublic sector2025 – ongoingOffice for National Statistics
LLM occupation classification for the Annual Survey of Hours and Earnings
ONS said its occupation-classification tool halved survey coding time
Problem
The Annual Survey of Hours and Earnings required employees to be assigned to occupation categories. Coding survey responses took staff time and affected data quality.
Approach
ONS moved an AI text-classification tool from prototype into live statistical production. The tool used large language models to assign occupations and was implemented for the 2025 results round.
Results
Takeaway
ONS reported both faster coding and more accurate assignments. It did not provide a numerical accuracy comparison in its economic-statistics plan.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
ONS reported that its occupation-classification tool improved assignments and halved coding time. It described this as its first use of large language models in ongoing statistical production.
Full details are in the original sources.