To create a usable library from a massive number of sources, a globally-respected institute of higher education needed to create metatag topics users can search. This caused a bottleneck as SMEs had to manually review all of them.

Primer’s solution was a Topic Text2Text model that automatically generates appropriate topic tags for any given article.

This case study details how Primer used their pre-trained models and AI training infrastructure to turn unstructured data into structured data and solve the problem.