The idea
What it is
The problem
A founder finishes the Tuesday newsletter and spends 20 minutes picking between 2 titles: "How I got my first 100 paying users" or "Some thoughts on growth this week". They go with their gut and find out the open rate the next afternoon, by then the whole list has already seen it.
A real A/B test needs enough subscribers to split. A 2-month-old newsletter does not have that.
The product
Paste 2 or 3 title options and pick the reader type, such as bootstrapped founders. A panel of AI personas built like that reader reacts to each title. You see how many would open, skip, click or get confused, and which words caused it.
Three things set it apart from JevTown, which shows 10,000 general personas reacting to any post.
- One audience and one format at a time, so the personas actually read like the people on your list
- Counts per title, so the choice becomes a number and not a paragraph of opinion
- Runs fast with a decision model such as Jev, so testing 3 titles fits in the minute before you send
The weekend version: newsletter titles only, 40 founder personas, and an open or skip count for each title.
Who pays
- Buyer
- A founder with a newsletter under 1,000 readers, in the minute before they send.
- Price
- Starting point: $9 a month per channel, under NowKnow's $15 Basic plan for AI persona feedback on ads and sites.
Why now
In September 2026 builders started using Jev, TypeSafe's model that makes fast structured decisions instead of writing paragraphs. JevTown built a town of 10,000 AI personas that react to any headline with it. The use case is still broad, which means every niche is open.
How to validate it
- Ask 10 newsletter writers in your niche how many title versions they wrote for their last issue. Pass: 7 or more wrote 2 or more.
- Score the titles of one founder's last 5 issues with 10 fixed personas and compare the ranking with their real open rates. Pass: the top-scored title is in the top 2 by real opens.
- Share a free beta in a solo-founder community. Pass: 5 or more people test again within 7 days.
- Ask 5 beta users to pay $9 for the next month. Pass: 3 or more pay.
- After 4 weeks, compare the open rates of users who sent the top-scored title with their own previous 4 weeks. Pass: at least half have a higher average open rate.