Case Study: Consensus App achieves state-of-the-art, highly accurate datasets in months with Datasaur

Consensus Creates State-of-the-Art Datasets in a Few Months

Consensus, a Boston-based tech startup building an evidence-based search engine, faced complex annotation challenges: custom multi-level labels (entity, sentence, multi-sentence), coordinating multiple annotators including PhD contributors, and needing oversight tools for dozens of annotators and projects. To train their ML model efficiently, Consensus partnered with Datasaur and adopted Datasaur’s annotation platform and services to meet these specific requirements.

Datasaur provided end-to-end support—custom demos, platform setup, training, and collaborative review workflows—enabling Consensus to produce thousands of high-quality annotated scientific papers and achieve highly accurate results within a few months. The successful first workflow and launch of a second workflow demonstrate measurable scale and speed improvements driven by Datasaur’s variable detection, sentence-level classification, and workflow management features.

Consensus

Eric Olson

Co-Founder & Chief Executive Officer