Yes, AI chatbots do cite press releases distributed through newswire services, sometimes within hours. That does not make a wire release a stand-in for a strong owned page. The available data points two ways: one vendor study found almost every release it tracked was cited at least once, while a broader citation study found press releases and newswire content made up only a tiny share of the citations it analyzed.
The practical answer: treat a wire release as a fast discovery signal that can support your owned content, not as the asset that builds your authority. Below you get the data, its limits, and how a release differs from a blog post on your own domain.
What the data says about press releases and AI citations
Two studies in the source set look at this question from different angles. Read them side by side, because neither tells the full story alone.
The newswire view: releases get cited, and fast
Notified, the company behind GlobeNewswire, analyzed releases distributed during one week in May 2026 across local, national and global circuits, with citation data from its partner Profound. It reports that 99.3% of the releases analyzed were cited by ChatGPT or Claude, with an average time to citation of eight hours.
Keep the limits in mind. This is one week of releases from one newswire, published by a company that sells press release distribution. It shows that releases were cited, not how often they appeared compared with other sources, or for which prompts.
The citation-share view: releases are a small slice
BuzzStream looked at the sources behind AI answers from the other direction. In its analysis, press releases accounted for only 0.32% of news citations, and newswire content made up 0.21% of the entire analysis. Brand newsrooms, where companies post their own releases and news, were counted separately: 3% of the whole analysis, and 18% of citations on ChatGPT compared with about 2% on Google’s platforms.
How both findings fit together
| Question | Notified (GlobeNewswire, May 2026) | BuzzStream |
|---|---|---|
| What was measured? | Whether distributed releases were cited by ChatGPT or Claude | Which source types appear across AI citations |
| Headline finding | 99.3% of analyzed releases were cited | Press releases were 0.32% of news citations |
| What it tells you | A release can get picked up | Releases were a small share of the citations analyzed |
A release can be cited somewhere and still be a minor source in the answers your buyers actually get. Both can be true at once.
Do AI chatbots treat a wire release as an independent credible source?
A press release is your own message, written by you and distributed by a paid service. The wire puts it on more domains. It does not turn it into independent coverage, and nothing in the source set shows that chatbots weigh it like an earned article by a journalist.
Being cited is not the same as being judged credible. Researchers who studied 10 AI-search platforms and 17,211 citation instances found that 8 of 10 platforms cited a fabricated concept within seven days after it was published on platforms they tended to cite. Citation selection, in other words, can follow where content sits rather than a careful check of who stands behind it.
So a citation of your release shows the chatbot found it useful for that answer. It does not prove your brand has gained the trust that independent mentions build. That earned side is covered in our guide to off-site distribution and AI citations, and the difference between mentions and links is explained in brand mentions vs backlinks for AI visibility.
Does wire syndication multiply your citation chances?
Not automatically. A wire pushes near-identical copies of your release onto many sites. More copies can mean more places for a crawler to find it, but identical copies compete with each other.
Microsoft’s Bing team explains that LLMs group near-duplicate URLs into a single cluster and then choose one page to represent the set, and that the chosen version may not be the one you intended. Bing notes that duplication can slow how quickly updates show up in AI-generated results.
For a wire release, that means a syndicating news site may be the URL that represents your announcement, not your own newsroom. Fifty copies do not equal fifty chances. They may equal one chance, with the credit landing on someone else’s domain. BuzzStream separately found that syndicated news sources such as MSN or Yahoo made up 6.2% of news citations in its analysis; that figure covers syndicated news in general, not wire release copies specifically.
Republishing your blog articles on partner sites raises a related set of choices. Those are handled in content syndication and AI citations.
Press release vs owned blog post: what differs
| Wire-distributed press release | Owned blog post | |
|---|---|---|
| Who controls the URL? | Mostly the wire and syndication sites | You |
| Copies online | Many near-identical versions | Primary version on your domain; copies may exist if it is republished |
| Lifespan | Tied to a moment: a launch, a result, a hire | Can answer an ongoing question and be updated |
| Role in the data above | Fast pickup; small share of citations | Notified reports owned content was twice as likely to be cited within 30 days when supported by a release |
That last row matters. In the same Notified study, owned content such as a blog or campaign page was twice as likely to be cited within the first 30 days when a press release supported it. The same limits apply: one newswire, one week, a vendor’s own report. Still, it points to the release as a companion to your owned page rather than a replacement.
Your own newsroom is the middle ground. BuzzStream found brand newsrooms accounted for 18% of ChatGPT citations in its data, compared with about 2% on Google’s platforms. Make sure your newsroom pages are indexed in the search engines your buyers’ chatbots may draw on, such as Google and Bing. Our Bing indexing checklist helps you check this.
Should a press release include quotable statistics?
The source set does not test statistics inside press releases. It does say something about structure: Notified reports that well-structured release content was the strongest predictor of AI visibility in its data, generating 2.6 times more citations on average.
Our advice, not a measured finding: write the release so a chatbot can lift a clear fact without guessing. That means:
- A first paragraph that states who did what, when and why it matters, in plain words.
- One concrete number or fact per claim, with its source and date.
- Your brand name written exactly as it appears on your own site.
- A link to the owned page that holds the full story, data and context.
If you publish original data, put the complete version on your own domain and let the release point to it. Our guide to the statistics tactic for AI-citable pages shows how to present numbers on that owned page.
How to use press releases without losing the credit
- Publish the owned page first. Make sure it is live and indexed before the release goes out.
- Post the release in your own newsroom. Give it a stable URL on your domain.
- Link from the release to the owned page. The release announces; the page explains.
- Keep the release for real news. A launch, a result or new data, not a rewritten blog post.
- Check which URL gets cited. Ask ChatGPT, Gemini and Perplexity the questions your announcement answers and note whose URL appears. The AI citation audit checklist gives you a repeatable routine.
A release can open the door. The pages behind it are where you hold the full story on your own domain. If you want a steady stream of citable owned content to support every announcement, see how Authora’s AI organic growth engine works.