I have a payroll software business and customer support is hectic. We started implementing AI workflows and had good early success, but quickly hit a ceiling due to a lack of good documentation. We looked at what we had and it was outdated and would take months to update with capacity that we didn't have. We thought about generating documents from source code, but there were too many levels of abstraction between our source code and the real experience of customers for this to be meaningful in a customer support context. But we were already spending a lot of time communicating real-world solutions to customers and getting direct feedback on what worked / didn't worked. We just weren't capturing and curating this knowledge. That's why we decided to build Knowledge Builder, to analyse resolved tickets and automatically create and enhance KB articles to improve documentation. It's made a huge difference at SuperPayroll for both human and AI agent performance.

Knowledge Builder for Freshdesk

From every ticket, a smarter knowledge base

From every ticket, a smarter knowledge base