Case Study: Timeform makes horse racing insights accessible with AI from Parallax

A Parallax Case Study

Preview of the Timeform Case Study

Timeform launches QuickCard with Parallax, achieving 95%+ F1 accuracy

Timeform, a part of the Flutter group, partnered with Parallax to create a new automated product called QuickCard. Their challenge was to translate their expert horse racing data and analysis into simple, jargon-free insights for casual fans, but they needed to ensure the AI-generated summaries were entirely automatic and highly accurate without any human input to avoid misleading information.

Parallax addressed this by working with Timeform's analysts to build a solution using Claude Sonnet 4 on AWS Bedrock, which restricted the AI to a curated set of reliable phrases. The resulting fully automated pipeline generates QuickCards for every race, achieving an F1 score of over 95% in evaluations. Since its launch, QuickCard has been successfully used at major race events, helping Timeform scale its AI capabilities to engage new audiences.


View this case study…

Parallax

30 Case Studies