Field notes · AI answer engines
What do AI answers actually cite?
We run real buying questions through ChatGPT, Perplexity, and the rest, record every citation, and publish the patterns we find. If your product never shows up in those answers, we can show you why.
Five prompts, one engine, a short written summary. No card, no call.
Diagnosis, not another dashboard
The default in this market is a monthly subscription that shows you a score and leaves the interpretation to you. We do the opposite: one engagement, a written report, raw answers attached, every finding with a fix you can hand to engineering.
| One-time diagnostic | Monthly monitoring tool | |
|---|---|---|
| You pay for | One engagement, scoped up front | A seat, billed until you cancel |
| You get | 20–30 real prompts run across the engines your buyers use, a written gap report with the raw answers, and a prioritized fix list | A dashboard of prompt rankings you interpret yourself |
| After it’s done | Nothing owed. Ongoing monitoring is optional, your call | Next invoice, same as this one |
This diagnostic is the service side of this site. We run it as CiteGap — the same work published here as field notes.
Latest notes
Best AI visibility tools for developer tools (2026): an honest tiering
Free checkers, $29 budget trackers, devtools specialists, enterprise platforms, and one-time alternatives, compared for developer tools companies. Prices checked October 21, 2026.
Field notesHow to check if ChatGPT recommends your devtool (manual method, one afternoon, no tools)
Build a 15 to 30 prompt test set, run it across ChatGPT, Perplexity, and Gemini, log mentions and citations in a spreadsheet, and compute your AI recommendation baseline. No tools required.
New notes when the data is ready. No schedule, no filler.