Philo2023

Philo was a greenfield product I led frontend development on, built as a Microsoft Outlook add-in running in both the browser and desktop versions of Outlook. It was conceived as a potential spinoff from Doma, applying machine learning capabilities developed internally to a broader document and email productivity problem.

The core idea was that professionals dealing with high volumes of contracts, title documents, and loan files could use Philo to get quick summaries of what those documents contained, surface actionable items from long email threads, and generate templated replies to speed up their responses. The ML services powering the summarization and document analysis were built by teammates and integrated into the frontend I built.

The add-in supported attachment previews, direct document uploads to internal order systems, and association of emails with specific orders, loan numbers, or addresses. Expanded search made it easy to find records across shared and individual inboxes without leaving Outlook.

Philo reached a late-stage internal beta and was presented to outside investors as a spinoff opportunity when Doma began winding down. The timing was difficult — funding was contracting across the board in 2023 — and the spinoff ultimately didn't come together. It was a strong proof of concept for what ML-assisted document workflows could look like in a high-volume professional context.