Try Pitch'd Free →
Pitch'd 2026 Audit

Why ChatGPT doesn't cite your releases.

Fourteen years of press releases. Eight AI-citation signals. Zero of 758 cleared the wire-ready threshold. Independent and free to download.

758
Releases audited
B2B press releases scored across 14 years, 2012 to 2026.
0
Cleared 80/100
Not one release scored high enough to be considered wire-ready for AI answer engines.
74
Highest score
The single best-performing release in the corpus. Six points below the threshold.
+22
Median editorial lift
Editorial pass delivered a median gain of +22 on Timeliness and +22 on Authority on the same release, draft to FINAL.
Pitch'd 2026 Audit
Why ChatGPT Doesn't Cite Your Releases
758
Releases · Fourteen years
Free download

The 758-release audit.

Pitch'd analyzed fourteen years of B2B press releases against eight AI-citation signals. Every headline, lead, quote, boilerplate, and closing paragraph was scored on the same rubric. The findings are honest and occasionally unflattering.

Length: 16 pages Corpus: 758 releases, 2012 to 2026 Format: PDF
Get the report → Try the Free Preview
Fill out the form below and download the report instantly. Free. No paywall.

Four findings that reframe the category.

The full report has more. These four are the ones that change the conversation.

1
The ceiling is real. Nobody hits it.
Zero releases in the corpus cleared the AI wire-ready threshold of eighty. The single highest score was seventy-four. This is not a sampling artifact. The empty space above seventy-four is a structural ceiling that the current press release conventions cannot clear.
2
The share fell by more than half.
In the Pre-AI era (2012 to 2019), 21.5 percent of FINAL releases scored "good for AI." In the AI era (2023 to 2026), that share dropped to 10.8 percent. AI standards moved upward. Writing practice did not move with them.
3
One editorial pass moves three signals.
Comparing draft to FINAL on the same release, editorial input delivered a median lift of +22 on Timeliness, +22 on Authority, and +14 on Specificity. The playbook works. Most releases just skip it.
4
Three factors drove the decline.
The eight signals do not decay evenly. Three specific factors account for most of the drop in AI citability across fourteen years. The full report names them and shows the trend lines.
Get the full report

The full audit, one form away.

Sixteen pages of data, charts, and the specific patterns that separate the top-scoring releases from the rest. Fill out the form and download instantly.

Your report is ready.

Sixteen pages of data, charts, and the patterns that separate top-scoring releases from the rest.

Download the report (PDF) ↓

Score your own release before the report even ships.

Every release in the corpus was scored by The Scout. See your overall score and weakest signal free in fifteen seconds. Full 8-signal breakdown and fixes unlock inside Pitch'd.

Try the Free Preview → See Pitch'd

How the audit works.

Every release was scored against the same eight signals that measure whether a press release can be cited by a large language model. The signals cover concrete details, quotable statements, editorial authority, originality, external citations, machine-parseable readability, time anchoring, and entity disambiguation.

Genre weighting is applied. A funding announcement is scored differently from a personnel move. The eight signals stay the same. Their weights adjust to reflect how each release type actually communicates.

Every score is reproducible. The scoring functions are the same code that runs inside The Scout preview and the Pitch'd builder audit. There is no opaque weighting, no vendor-tuned lift, no lookup table anyone can dispute.

The eight AI-citation signals
  • Specificity — concrete details AI can extract
  • Quotability — direct, attributable statements
  • Authority — structural credibility signals
  • Originality — distinctive versus generic PR
  • Citation Potential — external anchors AI can verify
  • Readability — how cleanly AI parses sentences
  • Timeliness — time anchoring for freshness grading
  • Entity Disambiguation — canonical name confidence