Legora reported using GPT-6 Astra to review 41 documents within minutes as part of a financial-review workflow. According to the description, the model identified all four errors that had been deliberately planted in the material and delivered a performance improvement of nearly 40%.
Why it matters
Document review in financial contexts can be time-consuming and error-prone. The stated results point to faster turnaround and accurate error detection in this specific test scenario.
Who should care
Teams handling financial-review workflows and organizations evaluating AI models for document analysis may find this example relevant.