News Content Loses Half Its ChatGPT Citations in 36 Days: What PR Teams Should Do About It (2026)

Promptwatch Data shows half of the news pages ChatGPT cites stop being cited within 36 days, while landing pages and docs last far longer. Here is how PR teams should rebalance coverage, owned pages, and measurement.

Key takeaways

  • In Promptwatch Data, half of the news pages ChatGPT cited on July 6, 2026 were no longer cited after 36 days. Almost a quarter were never cited again after day 0.
  • News is only 5.7% of ChatGPT citations (Sep 28 to Oct 4, 2026). Product pages (22.4%) and landing pages (16.8%) dominate.
  • Landing pages, docs, comparisons and listicles keep most of their citations after two months. Press coverage is a burst, not a base.
  • The fix is to pair every coverage push with durable owned or reference pages, then measure citations by engine and content type.
  • Treat the numbers as a snapshot. ChatGPT's source selection changed sharply in August 2026, and it can change again.

What the data actually says

PR teams have spent years treating a big placement as a lasting asset. In ChatGPT, it mostly isn't one.

Promptwatch took every classified page ChatGPT cited on July 6, 2026 (day 0) and followed those pages for 90 days. The report, What is the half-life of content in ChatGPT?, shows the share of those pages still cited on each later day. A content type hits its half-life where its line crosses 50%.

Favicon of Promptwatch

Promptwatch

Track and improve your AI search visibility
View more
Screenshot of Promptwatch website

The news results are blunt:

  • Almost a quarter of news pages were never cited again after day 0.
  • Half of news pages were gone after 36 days.
  • At day 60, ChatGPT still cited 41% of them.
  • Only 77% of news pages were cited again at any later point. For docs, it was 93%.

Compare that with other content types:

Content typeStill cited at 30 daysStill cited at 60 daysHalf-life
NewsBelow 50% by day 3641%36 days
Landing pages78%69%Likely near or beyond 90 days
Docs76%67%Likely near or beyond 90 days
Comparisons72%63%Likely near or beyond 90 days
Listiclesn/a60%Final week of the window
Product pages, reviews, how-tosn/aAbout 55%74 to 77 days

One caveat from the report matters. Lines drop faster in the final week because fewer days remain for a page to be cited again. Landing pages, docs, comparisons and listicles only cross 50% in that last week, so their real half-life is probably close to or beyond 90 days. News is the only type that clearly decays in a matter of weeks.

News is a small slice of what ChatGPT cites

The decay rate would matter less if news were a big share of citations. It isn't. In Promptwatch's ChatGPT citation share by content type (Sep 28 to Oct 4, 2026):

Content typeShare of ChatGPT citations
Product pages22.4%
Landing pages16.8%
Listicles13.6%
How-to guides8.7%
Documentation7.5%
General articles6.1%
Comparisons5.9%
News5.7%
FAQ pages3.3%
Reviews1.7%

Product and landing pages together make up 39% of citations. Editorial formats together make up 43%. So journalism is part of the picture, but news specifically is a thin, fast-fading layer of it.

The prompt type shifts the mix:

  • Commercial prompts: product pages 26.0%, landing pages 20.6%, listicles 15.0%, comparisons 7.9%.
  • Informational prompts: product pages 19.1%, landing pages 13.9%, listicles 13.6%, how-tos 10.7%, docs 8.5%.
  • Branded prompts: product pages 23.3%, landing pages 22.8%, reviews 5.6%, FAQ 4.8%, social posts 3.0%, press releases 2.1%.

Press releases earn a larger share on branded prompts (2.1%) than across all prompts, though the report doesn't give the all-prompt figure. When someone asks ChatGPT about your brand by name, your own pages and reviews carry the answer. Coverage is a supporting actor.

Why this does not make PR irrelevant

It would be a mistake to read this as "stop pitching journalists." Two reasons.

First, earned media still feeds AI answers heavily. Muck Rack's May 2026 report analyzed more than 25 million links cited by ChatGPT, Claude and Gemini on consumer prompts and found earned and third-party sources cited far more than paid ones. Muck Rack sells a tool for tracking AI citations, so weigh its commercial interest. It also defines earned media broadly: journalism, academic research, government sources, encyclopedic sites and third-party corporate content. That 84% figure is not a count of news articles.

Second, the same research found that the journalists most pitched on Muck Rack's platform and the journalists most cited by AI for a given brand overlapped by only 2% on average. A media list built for traditional reach can miss the outlets AI engines actually read.

So the issue isn't that coverage is useless. It's that a news article is a short-lived citation, and plenty of PR plans stop there.

A reminder that ChatGPT's source mix moves

The decay data describes a July cohort. Source selection itself can shift quickly, and PR teams should keep that in mind before locking in a strategy.

In mid-August 2026, reddit.com's share of ChatGPT citations collapsed from roughly 4% to 0.5% on August 14. Promptwatch Data on Reddit citations dropping in ChatGPT tracks it daily, and AI Overviews and AI Mode declined more gradually. Around the same time, Promptwatch recorded ChatGPT Search using the site: operator at scale from August 8, jumping from about 0.4% to about 17% of fanout queries (report).

Otterly.AI studied 16 brand reports across unrelated industries and found Reddit citations from ChatGPT fell by at least 73%, while ChatGPT's total citation volume rose 3.5%. Other engines kept citing the same threads. Their read: the gap was filled by reference sources and official pages rather than other forums.

Otterly.AI study on ChatGPT Reddit citations and what replaced them

WPP Media was careful about this too: the cause isn't confirmed, and it is too early to call it a permanent change. Their advice was to understand where visibility exists across engines, which sources are cited, and how patterns evolve.

I think that's the right posture. Nobody outside OpenAI knows the retrieval design. What you can say is that a shift toward canonical and first-party sources fits with the news decay data. When the engine leans on official pages, durable owned content has an edge over a one-week news cycle.

What PR teams should do

1. Treat coverage as a burst and plan for the decay

Promptwatch's own recommendation is to use news and press coverage for short bursts of visibility. Plan the timing. If a launch story lands in week one, assume its ChatGPT citation value is mostly gone by week six. Schedule the next beat before the first one fades, rather than treating one hit as a permanent win.

2. Give every story a durable home on your own site

For each major announcement, publish an owned page that holds the facts: a landing page, a documentation page, a comparison, an FAQ. The data says these stay cited longest. Landing pages kept 78% of citations at 30 days and 69% at 60 days. Docs were close behind.

Press coverage can then point back to this page, and the page stays in the answer after the article has dropped out.

3. Keep the facts identical everywhere

AI engines repeat what they can verify. Letting facts differ between your website, profiles and press materials is one of the more common avoidable mistakes. Publish claims with a named source and a date. A release with verifiable numbers can itself be cited, so treat it as a source document, not a formality.

4. Invest in comparisons and listicles

Comparisons keep 63% of citations at 60 days and listicles 60%. Listicles are also 13.6% of all ChatGPT citations. If you only get cited in articles written by others, consider owning a fair, well-sourced comparison or roundup in your category, and earning inclusion in third-party ones.

5. Keep product pages current

Product pages are the biggest single citation type at 22.4%, but they decay about as fast as how-tos, reaching half-life after 74 to 77 days. Stale pricing or feature claims will keep getting repeated until someone fixes them. Assign an owner and a refresh cadence.

6. Rebuild your media list around what AI cites

Given the 2% overlap between most-pitched and most-cited journalists, check which outlets show up in AI answers for your core topics and pitch those too. This is a measurement task, not a guess.

7. Watch reviews on branded prompts

Reviews are only 1.7% of citations overall but 5.6% on branded prompts. If someone asks about you by name, review pages show up more than you'd expect. Make sure the ones you control are accurate.

How to measure it

Coverage counts won't tell you whether any of this works. You need to see the answers.

A useful measurement setup tracks:

  • Which of your pages and which third-party articles get cited for your priority prompts.
  • The content type of each citation, so you can see whether you depend on news.
  • How long a cited page keeps getting cited after publication.
  • Results by engine, since ChatGPT, AI Overviews and Perplexity behave differently. Promptwatch's social citations report found ChatGPT drew 0.46% of citations from social platforms against 12.2% for Google AI Overviews (Sep 28 to Oct 4, 2026).
  • The denominator. A drop in one domain means little unless you know total citation volume didn't also fall.

That last point is the lesson of the Reddit episode: carry a control next to every metric.

Tool options

Promptwatch is the one I'd start with for this, because it classifies citations by content type, shows citation trends over time with first cited date and decay half-life per page, and also covers crawler logs and AI traffic. It also has a content gap analysis and Content Agent for publishing the durable pages once you know what's missing. You can start with the free AI visibility report or the 7-day free trial.

Other monitoring tools worth a look depending on budget and focus:

Favicon of Otterly.AI

Otterly.AI

Affordable AI brand visibility monitoring
View more
Screenshot of Otterly.AI website
Favicon of Peec AI

Peec AI

AI visibility tracking with smart suggestions
View more
Screenshot of Peec AI website
Favicon of Profound

Profound

Enterprise AI search visibility and analytics
View more
Screenshot of Profound website
Favicon of Scrunch AI

Scrunch AI

AI search monitoring for brands and agencies
View more
Screenshot of Scrunch AI website

Otterly.AI and Peec AI are lighter prompt trackers. Profound and Scrunch are stronger on enterprise monitoring. For a broader list, the GEO software directory at bestgeosoftware.com is a good place to compare options.

If outreach is the bottleneck, Muck Rack-style media databases are one route, and tools like [tool:respona] can help with the pitching workflow.

A simple 90-day plan

TimeframeAction
Week 1Pull your priority prompts, see which pages and outlets get cited, and note the content type of each
Weeks 2 to 3Audit owned pages for accuracy: landing pages, docs, pricing, FAQs. Fix inconsistencies
Weeks 3 to 6Publish a durable page for each major story, plus one comparison or roundup per core topic
Weeks 4 to 8Pitch the outlets that AI actually cites, not just the biggest ones
Weeks 8 to 12Re-measure by engine and content type. Refresh decaying pages. Plan the next news beat before the last one fades

What not to do

  • Don't pitch only the outlets with the biggest reach without checking which ones AI answers cite.
  • Don't judge success on coverage counts alone.
  • Don't rebuild strategy around a four-day window. The Reddit shift is a reminder that short windows mislead.
  • Don't assume the 36-day figure is a law. It's one cohort of pages from one start date, with the end-of-window caveat noted above.

The short version

News gets cited briefly and in small amounts. Landing pages, docs, comparisons and FAQs get cited steadily. A PR plan that ends at the placement hands ChatGPT a story that expires in about five weeks. A plan that ends at an accurate, durable owned page gives it something to keep citing. For more research on how AI answers are built, the Promptwatch Data hub is updated regularly.

Share:

AI Search Visibility Tools

© 2026 AI Search Visibility Tools · The best AI search visibility tools compared · RSS

AI Search Visibility Tools is an affiliate review site. When you click links to vendors or buy through links on our site, we may earn an affiliate commission at no extra cost to you.

AI Search Visibility Tools is a review website based on user reviews on Reddit and G2, and on publicly available information. We keep everything as up to date as possible, but pricing and features can change. Always confirm the details with the vendor before purchasing.