Profound, an AI visibility platform, reported a median of 6.81 days between a page's publication and its first citation by ChatGPT or Claude. That figure comes from newly published marketing pages that were eventually cited, so it describes pages that succeeded, not every page. A press placement can appear in AI answers within days, but no agency can guarantee when or whether it will.
Last updated: October 24, 2026
In this guide
- What does Profound's research say about time to first citation?
- What does the 6.81-day median leave out?
- Why does citation timing matter for PR teams?
- How does a placement reach an AI answer?
- What affects how fast a placement gets cited?
- How should a PR team track time to citation?
- How should you read a citation timeline report?
- What should an agency never promise about timing?
- Frequently asked questions
What does Profound's research say about time to first citation?
Profound's reported median time to first citation is 6.81 days for ChatGPT and Claude combined. The figure comes from a LinkedIn post by Josh Blyskal, a member of Profound's founding team, as described in a newsletter analysis of that post.
| Figure | What it measures |
|---|---|
| 6.81 days | Median days from publication to first citation by ChatGPT or Claude |
| 37 days | Point by which 90% of the sampled pages had their first citation, as reported |
| About 900 pages | Newly published marketing pages in the sample |
| 60 days | Observation window in Profound's logs |
What does the 6.81-day median leave out?
The 6.81-day median leaves out every page that was never cited, because the sample only included pages that ChatGPT or Claude eventually cited. A page that no engine picks up does not appear in the number.
The figure also blends two engines and covers marketing pages, not press placements specifically. Treat it as a benchmark for one dataset, not a promise for any single placement.
Why does citation timing matter for PR teams?
Citation timing matters because it sets the reporting cadence and the expectations a client holds. A team that expects a citation the day a story runs will judge a placement a failure before an engine has had time to find it. A team that waits months with no check will miss early signals.
Timing also affects how a team compares placements. Two stories published the same week can reach AI answers at different speeds, depending on the outlet, the topic and the questions people ask.
How does a placement reach an AI answer?
A placement reaches an AI answer in four steps: the article is published, a search index or crawler finds it, an engine retrieves it when answering a question, and the engine chooses to cite it. A delay at any step delays the citation.
The retrieval step depends on a question being asked. An engine only cites a page when a user's prompt leads it to search for something the page answers. That is why a strong placement on a rarely asked topic can sit uncited for a long time.
The selection step depends on the page itself. Articles that state the brand's name, the key facts and the answer to a clear question in plain sentences give an engine something it can quote directly.
What affects how fast a placement gets cited?
Four factors affect how fast a placement gets cited: whether the page is indexed, whether the engine retrieves live web results for the question, how clearly the page answers that question, and whether anyone asks the engine about the topic at all.
These factors are general mechanisms, not measured effects from Profound's data. Outlet choice and how plainly the article states the brand's key facts also shape what an engine can quote.
How should a PR team track time to citation?
A PR team should record the publication date of each placement, then check a fixed set of prompts on a schedule and log the first date an engine cites the page.
- Record the publication date and URL of each placement.
- Write the prompts a buyer would ask that the placement could answer.
- Run those prompts weekly in ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews.
- Log the first date each engine cites the page, and the engine.
- Keep placements that were never cited in the same log, so the results are not skewed toward wins.
The template below shows the columns to keep. It is a blank format, not real results.
| Placement and outlet | Publication date | Prompts tracked | Engine | First cited date | Days to citation |
|---|---|---|---|---|---|
| Story name, outlet | Date | Prompt list | Engine name | Date or "not yet" | Calculated |
How should you read a citation timeline report?
A citation timeline report should be read against its sample. Ask which pages were counted, which engines, and over how long. A median from pages that were all eventually cited will look faster than a result that includes pages that never were.
Look for the same fields in any report an agency gives you: how many placements were tracked, how many were cited, how long each took, and which engine cited each one. A report that shows only the fast wins hides the uncited placements, which are part of the picture.
What should an agency never promise about timing?
An agency should never promise a citation date, a citation count or a specific engine. AI engines choose and change their own sources, and a published median describes a sample, not a schedule. An agency controls pitching, writing and page structure, and it can report results against tracked prompts. For how this work is delivered, see the GEO optimization practice page.
For the wider picture, read how earned media shapes AI search visibility, and use the 10 checks and package scope guide when you compare agencies.
Frequently asked questions
How long does a press placement take to appear in AI answers?
Profound's reported median is 6.81 days for pages that were eventually cited by ChatGPT or Claude. Individual placements vary, and some are never cited.
Is the 6.81-day figure a guarantee?
No. It describes a sample of about 900 pages that were eventually cited, not every placement.
Does every AI engine cite new pages at the same speed?
No. Profound's figure combines ChatGPT and Claude, and other engines were not part of that number.
Why was my placement not cited?
The page may not be indexed, the engine may not have been asked a question it answers, or another source may have fit better. Track the same prompts over time before drawing conclusions.
Can an agency speed up citations?
An agency can improve conditions, such as outlet choice and clear page structure, but it cannot control when an engine cites a page.
