Currently, trigram similarity scores are calculated for chat cases and available via CSV export of QA review data, but they are not visible within the Assembled UI itself.

Customers with high chat volumes struggle to access trigram similarity scores because exports are slow and require very narrow filters (single-day date ranges) to complete successfully

The trigram similarity score is a primary metric for measuring co-pilot accuracy, especially for chat-focused teams

The current CSV export workaround is not sustainable for ongoing monitoring and reporting

Add the trigram similarity score as a visible field in the QA Review interface, similar to how other metrics are displayed.