Hey PH, Tyler here, cofounder of Praxis. We took leave from Penn M&T and Princeton three weeks ago, and have been on the road talking to campaign managers and iterating on our product every day since. Here's what we kept running into: every election cycle, volunteers have millions of conversations with voters. They figure out what actually moves people, which issues, which framing, which tone. Then the campaign ends and all of it disappears. Next cycle starts from scratch. We started asking: What if you could capture that? Praxis enriches a campaign's voter file and generates a personalized script for every call. A voter who cares about climate hears about the candidate's clean energy plan. A renter whose building was just rezoned hears about her housing proposal. When the volunteer finishes the call, the response is auto-classified: no clipboards, no survey forms, no data entry. Every conversation feeds a persuasion dataset: What messages move which people, and why. We also built P2P texting and digital ad generation from field insights, a call at 10am becomes a targeted ad by 2pm. Campaigns are the wedge. The dataset is what we're really building. Would love feedback from anyone in politics, campaigns, or civic tech!

Praxis

AI voter contact that learns what persuades

AI voter contact that learns what persuades