用 Jev 给非营利机构捐赠人记录打分与邮件分流

Mahesh Murthy @mvmurthy原推发布于 2026.09.22

在 X 上打开原推

The new AI model Jev by @typesafeai is all over my X timeline with some really interesting usecases. It got me thinking about nonprofit use cases of Jev. Say your CRM has 1,000, 10,000, or even a million donor records. Jev could score each record for something specific like “likely to respond to personal outreach.” Then it's easy to just rank those scores and give your fundraising team 10 people worth looking at today. At TypeSafe’s published price of $0.042 per million input tokens, this could be cheap enough to run every day instead of doing one big segmentation exercise every few months. Another example: Someone emails asking about a program. Jev could score whether the message probably needs: - a phone call - a personal email - an automated reply - spam review And it would use relevant context from your CRM when producing those probabilities. For ranking and routing work with clear rules and a human override, I think there are a lot of possibilities here. Still thinking this through.