video1.0<iframe src="https://www.loom.com/embed/2ab27874f59c46febffe4f3335c7d3fb" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/2ab27874f59c46febffe4f3335c7d3fb-72559d76abf4e65f.gif300.117This Loom showcases an agent built in Profound to improve subscription ARR deal source attribution when HubSpot SDR notes are missing. The author explains that over 90% of recurring subscription ARR revenue previously had no deal source, often because SDRs only added minimal notes like “heard from LinkedIn” or “heard from X,” which disrupted marketing attribution. The agent uses a Python script reading HubSpot notes and then applies criteria across sources such as LinkedIn, X, Reddit, webinar, and events to recommend an attribution and post a Slack message to the deal. Two examples are demonstrated, including one ambiguous case where the owner could not recall how the relationship began after a first meeting, resulting in unresolved status and no recommended source.