United KingdomPublic sector2025 – ongoingUK Department for Business and Trade
Business Intelligence Subject Tagger
DBT used overnight feedback tagging with training-time precision as low as 0.34
Problem
Staff had to read business feedback in full to identify the subjects companies were raising. Reviewing new material could take hours.
Approach
DBT built supervised classifiers for predefined subject tags. The system labelled feedback from its customer relationship management platform overnight. Staff subsequently read the feedback in full before using it in briefings.
Results
Takeaway
DBT accepted incorrect tags to reduce missed topics. Staff review remained necessary because some tags had low precision.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Internal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
DBT documented a production tagging system and mandatory subsequent reading by staff. It reported precision ranging from 0.34 to 0.84 across deployed tag models, assessed during training.
Full details are in the original sources.
