The emergence of AI has turned everything upside down in digital marketing.
One of the most important challenges we have is to be able to measure the traffic that is coming to websites from generative AIs such as ChatGPT, Perplexity, Gemini, Copilot, Claude, etc.
We are already aware of the impact that the arrival of AI Overviews, or even AI Mode, has on organic traffic.
Now we seek to generate traffic also in those chats or search engines with AI, to have the largest number of possible channels, and thus counteract the loss that is being suffered in Google.
Is it Possible to Measure the Traffic that Comes from AI Overviews or AI Mode?
Unfortunately, no. At least not yet.
Knowing how many clicks or traffic you get from results in which you appear in Google’s AI summaries? It is not possible.
In the future, at least in AI Mode, we may have more data because Google has stated so.
They will remove the “noreferrer” and we will be able to have data in Google Search Console.
But right now there is no way to have that data.
Another thing is to sense that certain keywords are appearing in AI Overviews after analyzing click, impression, or CTR data.
For now, we settle for being able to see the traffic that comes from other search engines with AI or Chats.
How to Analyze the Traffic that Comes to your Website from ChatGPT, Gemini, Copilot and Others
The best way to do this is with a tool like Google Analytics.
Traffic from these places is tagged as a reference and is therefore included in the “Referral” channel or also as a referral in Source.
In these reports we can see the sessions, users, sales, income, conversions… that are generated from these channels.
I am going to explain how to get this data yourself from Google Analytics and Looker Studio, in a template like this.
👉Check out our selection of Looker Studio templates
From Google Analytics 4
To see the traffic that comes from AIs you can go to the Acquisition > Traffic acquisition report:
In the table you will see everything sorted by channels, but we want to see it by Source/medium.
Therefore, use this drop-down and search for source/medium:
Now you will see the different sources and mediums that generate traffic, among which you may find ChatGPT and others, but it is unlikely that you will find the first ones.
Use this search engine and put “referral”:
This way you will only see the source/medium that include “referral” as a medium.
And if you want to go into more detail, search:
- chatgpt
- gemini
- perplexity
- copilot
- claude
- mistral
- etc
This way each of the ones you are looking for will appear and you will be able to know how many sessions, users, sales, conversions they have generated.
This is the fastest way to see what each AI or LLM generates on your website, but it is not the best.
Ideally, you would have a report where you could see it whenever you want, when you enter, without doing anything else.
Even having more graphs that give you more information about the evolution of this type of traffic.
To cover this objective we are going to use Looker Studio.
This free Google tool that is so useful for creating personalized dashboards with the graphs and data we want.
Analyze Traffic from AIs with Looker Studio
In the case of Looker Studio we are going to use Google Analytics 4 as a data source, and we will follow the same logic of filtering by source/medium.
We can create several graphs that give us very valuable information.
I am going to show you how we can create some of the graphs that I include in this report:
Filtered Metrics:
We add result cards to the report with metrics such as Sessions, Total users, Views, Bounce rate, Conversions, Transactions…
If we do not filter those metrics, they will give global data, from all channels and sources.
We are going to create the filter that we will apply to these graphs and others that we will use later.
We can create the filters in the configuration of any graph or from Resources, in the top bar.
The filter configuration will be as follows:
As you can see, we are using a regular expression, to tell it that we want the sources that contain what we include in the box:
chatgpt|mistral|gemini|perplexity|copilotIn the event that you want to add more, you can do so by separating with “|” the new ones you add.
Once you save this filter, you can apply it to all the graphs and metrics you have in the report so that they are filtered by traffic from those places.
Table with the Distribution of Traffic from each AI
Another graph that we can use is this table:
With it we can see how traffic and other metrics are distributed by each AI.
For this, we use a table with metrics such as Sessions, Users, Conversions, Transactions… and as Dimension “Source”.
We assign the filter that we have just created to this table and we will have it filtered, showing only the sources that correspond to what we want.
Time Series with the Evolution of Traffic from Sources with AI
With this graph we can see the evolution in the last months, 12 for example, of the traffic that comes from AIs.
We can use a metric such as Sessions or Users, and as Dimension Date, since it is a time series.
We are going to manually assign a period of the last 12 months:
And to be able to see 1 month at each point, configure the Dimension in this way, with Month and Year.
And to finish, we assign the filter that we have created before.
Most Visited Pages from AI Sources
Another important aspect that we can analyze is the traffic that arrives for each landing page, and even have the breakdown for each page and the AI that has derived the visit.
For this, create a table and assign the filter that we have already used.
As Dimensions you can put Landing page and Source. Being filtered, only the sources that you have indicated in the filter will appear.
Then you can include the metrics that are most useful according to your case.
These are some of the ways you have to analyze the traffic that comes to your website through AI.
Surely if you start analyzing you will get some surprise.
Alex Serrano
12 años en marketing digital. Creador de Chartud y especialista en Looker Studio. Consultor SEO desde hace 8 años. Creador y divulgador de contenido en diferentes formatos como 300Segundos (newsletter) o SEOdesdeCero (podcast). También cofundador de RankPulse.app, herramienta de SEO Local.
