Skip to main content

Goal and Result of Fine-Tuning Requirements

Optimizing AI-based contract review through manual annotation

Updated over a month ago

The Goal: Significant Increase in Accuracy

By regularly and carefully reviewing these test sets, you can significantly improve two essential areas:

  1. The quality of your Playbook (the stored review requirements).

  2. The performance of the AI in the automated review of documents.

The Result of Fine-Tuning:

Through proper training (two essential principles: rigor and coherence) and the use of high-quality test sets, it is possible to increase the standard AI accuracy from 80–90% to over 95%.

Did this answer your question?