How we build our agent profiles
A plain-English look at where each profile comes from, how we verify it, what we deliberately leave out, and how a human reviews every page before it publishes.
We think you should know how the page you’re reading was made. Here’s the whole process, start to finish — including where AI helps and where a person has to sign off. The short version: we gather public information, synthesize it into an original profile with AI assistance, and a human reviews every page before it publishes.
1. Gather what’s public
For each agent we pull together the kinds of information a thorough writer would find on their own: the agent’s own stated submission guidelines and wish list, their agency bio and client list, public interviews and podcasts, public announcements of books they’ve sold, and their own timely social posts about opening or closing to queries. We only use information that’s already public. The full disclosure is on our sources.
2. Synthesize — never copy
We don’t republish source text. We fuse those fragments into one original summary written in our own words. This is where software, including AI, does the heavy lifting: assembling the material and drafting a first version at a scale one person couldn’t manage by hand. The goal isn’t just to restate facts but to surface the useful pattern — for instance, what an agent’s sold books quietly have in common, which often says more about their taste than any single wishlist line.
3. Apply hard rules
The drafting process runs against a set of non-negotiable rules, because a fluent summary that drops a qualifier can send you to the wrong agent. Among them:
- Preserve meaning-changing qualifiers — “picture books from author-illustrators only,” “middle grade only from authors who also write YA,” “as a thread, not a main genre.”
- Never fabricate a date, quote, sale, or credential. If it isn’t verifiable, it doesn’t appear as fact.
- Date volatile facts like query status, and flag them when they age.
- Never infer personal details from a name — pronouns come only from an agent’s own public bio.
4. A human reviews every page
Before anything publishes, a person reads it: checking accuracy against the source material, cutting anything that can’t be supported, making sure the qualifiers survived, and fixing tone. Pages that trip our quality checks are held back, not published. AI assists; a human decides. There is always someone accountable for what’s on the page — see our editorial policy.
5. Publish, date, and keep current
Each profile ships with a date so you can see how fresh it is. We refresh status and signals on a regular cadence; a newer, more authoritative reading always replaces an older one. When we’re not confident something is current, we say so rather than implying it’s live. The criteria and labels behind it all are documented on how we evaluate agents.
6. Fix it fast when we’re wrong
Summarizing public information at scale means we sometimes get a detail wrong or fall behind a change. When that happens we want to know. Corrections — especially from the agents themselves — are a priority; just tell uswhat’s off and we’ll fix it.
And one thing we can’t do for you
No summary, however careful, replaces an agent’s own current guidelines. Whatever a profile says, confirm the submission window and instructions on the agent’s own page before you query. We build these profiles to get you to that page faster and better informed — not to be the last word.