Guide · ChatGPT traffic

A new site suddenly gets a wave of visitors from ChatGPT: what does it look like? Four days at Gamma AI Slide

Updated 2026-11-13 · kagentx

Short answer

A steep four-day slope: Gamma AI Slide began receiving visitors with chatgpt.com as the source on 2026-11-09, then 42, 62 and 65 sessions a day; in the last 7 days there were 244 visitors (19 in the previous 7), and by session source chatgpt.com 173, direct 83, google organic 19. In the same period 108 sign-ups, 70 generations, 0 payments. 77% of these visitors engaged, but we do not know what question ChatGPT was answering when it gave the link, and the dashboard has no such data, so we do not guess at causes.

When did a wave of ChatGPT visitors start and how many came each day?

Gamma AI Slide started getting visitors with chatgpt.com as the source on 2026-11-09, and then 42, 62 and 65 sessions a day over the next three days. Before that this source never appeared in our data dashboard (per-source data there starts on 10-15).

Date (2026)chatgpt.com sessionsSite visitorsSite sessionsSign-upsGenerate
11-0802200
11-094141430
11-104259733122
11-116278943225
11-1265851054223

Source: local dashboard, read 2026-11-13, GA4 (sessions, visitors) and site statistics (sign-ups, generations). The day of 11-13 was not over at the time of reading (8 chatgpt.com sessions) and is left out. Visitors are the daily visitor counts added up.

How many visitors, sign-ups and payments did the week bring?

In the last 7 days (11-06 to 11-12) there were 244 visitors, against 19 in the previous 7 days (10-30 to 11-05); in the same period 108 sign-ups, 70 generations and 0 paid orders. By session source, chatgpt.com has 173, direct 83 and google organic 19.

Session source (11-06 to 11-12)SessionsEngaged sessionsShare
chatgpt.com / ai-assistant17313477%
direct832024%
google / organic191158%

Engagement uses GA4's “engaged sessions” (longer than 10 seconds, or a key event, or 2 or more page views; any one of the three). 77% of chatgpt.com visitors engaged, higher than both google organic and direct; the same comparison on MochiStamp is in Does traffic from ChatGPT convert better than traffic from Google?. The sample is one site and one week, and engagement is not the same as value.

How far did these visitors get on the site?

From signing in to generating a deck to viewing pricing, every step has fewer, and payments are 0. Below are the GA4 events for 11-06 to 11-12: the first column is the number of events, the second the number of people who triggered the event.

Step (event)CountPeople
login_open (opened login)199123
login_click (clicked login)152114
sign_up (signed up)10193
outline_start (started an outline)11874
deck_generated (generated a deck)5648
pricing_open (opened pricing)4626
credit_wall (hit the credit limit)4025

Events are counts and one person can trigger several, so do not divide one row by the previous one as a conversion rate; the people column is closer to “how many people reached this step”: 93 people signed up, 48 generated a deck, 25 hit the credit limit, and nobody paid.

Site statistics count 108 sign-ups and GA4 has 101 sign_up events; the two figures differ a little and we have not checked why one by one.

Why did these visitors come? What was ChatGPT answering at the time?

We do not know. The dashboard has neither the content of ChatGPT's answers nor any parameters in the links; we do not guess at causes and do not put it down to “one page being recommended by ChatGPT”. Here are a few things we can see in our data, and their limits:

  • Landing pages: in the sessions of 11-09 to 11-13, /fa (the Persian home page) 97, the home page 67, /pdf-to-slides 17, /ar 15, /edit-powerpoint-with-ai 14. The dashboard does not cross landing pages with source, so we do not know which pages the chatgpt.com sessions landed on.
  • Countries: among the visitors of 11-10 to 11-12, Iran 62, Azerbaijan 31, the United States 22, China 14, India 11. Again not crossed with source, so we cannot say these people all came from ChatGPT.

There are many possible explanations: an answer in some conversation cited a page on the site, a user pasted the link into a conversation, a share was forwarded… none has data behind it, so we draw no conclusion. How Persian and Arabic pages perform in search is in Arabic and Persian pages in search and AI answers: what one small site saw.

If your site also suddenly gets a wave, what should you check first?

  1. 1. First confirm the source is real

    In GA4 acquisition, check that the source is chatgpt.com and the channel is AI Assistant. Setup steps are in How do you see ChatGPT traffic in GA4?.

  2. 2. Check the start date against what you did those days

    Our site's started on 11-09. Put the dates of your deploys, posts and external mentions in one table before talking about “possible causes”.

  3. 3. Look at landing pages and countries separately

    If your data can cross landing pages with source, look at which pages the chatgpt.com sessions landed on; our dashboard cannot, and that is our gap.

  4. 4. See what is left at each step, not only sessions

    Walk step by step from sign-in to payment, like the table above. In our week: 108 sign-ups, 70 generations, 0 payments.

  5. 5. Judge whether it lasts after a few weeks

    We have only 4 days of data and do not know whether it will last or fade. How to measure over time: How do you measure AI visibility? Four metrics and a table you can follow; how to make ChatGPT more likely to recommend you: How do I get ChatGPT to recommend my product?.

What are the limits of these numbers?

  • One site and a 4-day peak; the data of 11-13 is incomplete and not included.
  • In the dashboard, source, landing page and country are not crossed, so we do not know who the chatgpt.com sessions are or where they landed.
  • We do not know why ChatGPT gave these links and draw no causal conclusion.
  • Engagement is not value, and payments are 0; the sample is too small to extrapolate a conversion rate.

The direct and (not set) shares at four sites over a comparable week are in What are (not set) and direct traffic in GA4? Traffic source shares at four sites over the last 7 days.

FAQ

Is it normal for a new site to suddenly get many visitors from ChatGPT?

One of our sites went through it: Gamma AI Slide from 11-09, with 42 to 65 sessions a day for three days. With one site and 4 days we cannot say “normal” or “usually like this”.

Do ChatGPT visitors sign up and pay?

On our site in the last 7 days: 108 sign-ups, 70 generations, 0 payments; of the people, 93 signed up and 25 hit the credit limit. The sample is small, and 0 payments does not mean it will stay 0.

How do you find out why ChatGPT linked to my site?

Usually you cannot from site data alone. You can ask ChatGPT your buyers' questions a few times yourself and see whom it recommends and which pages it cites; that is what an evaluation does, and it can be done by hand.

Are ChatGPT visitors better quality than Google's?

In our data engagement is higher (77% against 58%), but that is engagement, not payment; the MochiStamp comparison is in another of our guides.

Will this wave last?

We do not know; we have only 4 days of data.

Sources

kagentx

Do AI answers
mention you?

An evaluation asks ChatGPT, Perplexity and Claude your buyers' questions (with web search on) and records who they recommend, which pages they cite and whether you are among them. Basic $1, results usually in 2–3 minutes.