How ProPublica appears in ChatGPT, Perplexity, Claude, and Google AI Overviews
A free AI-visibility snapshot for ProPublica (propublica.org) — a nonprofit newsroom based in New York, NY.
Overall AI visibility
propublica.org
65/100
Grade C
Across the four AI engines we audit (ChatGPT, Perplexity, Claude, and Google AI Overviews), ProPublica is inconsistently cited by the engines we tested, with an overall 65/100 score. The largest open lift is on the engines below where ProPublica's articles are missing structured signals other nonprofit newsroom sites of the same size already publish.
ChatGPT
62
Crawler: GPTBot
GPTBot is allowed but article-level structured data is thin.
Perplexity
43
Crawler: PerplexityBot
PerplexityBot blocked or paywalled before the lede.
Claude
87
Crawler: ClaudeBot
ClaudeBot crawls clean HTML; Anthropic citations appear for evergreen pieces.
Google AI Overviews
68
Crawler: Google-Extended
AIO surfaces the site occasionally; missing FAQ + speakable hurts.
What ProPublica can ship this week
- Add an XML sitemap entry for every article within an hour of publish so crawl latency for AI bots stays under 60 minutes.
- Add a public reporter bio page for every byline with `Person` schema, beats, and a sample of recent articles.
- Allow `GPTBot`, `PerplexityBot`, `ClaudeBot`, `anthropic-ai`, `Google-Extended`, and `Applebot-Extended` in `https://propublica.org/robots.txt`.
- Cross-link each article from at least two evergreen topic pages so AI engines can find it after the homepage rotates it off.
- Publish a `/llms.txt` at the root that maps your beats, reporter bios, and the canonical URL for each topic.
If you work at ProPublica — or any other nonprofit newsroom — the same audit, but live and continuous on every article you publish, is what Mani does.
Start a 7-day free trial →
$20/mo after trial. No card to start. See integrations.