The biggest referrer to your content might never show up in your Analytics platform. Let that sink in for a minute. It’s a problem for businesses who need visibility in AI search, and let’s face it, that’s everybody. Right now, if ChatGPT, Claude or other AI assistants quotes your blog post in a live answer, the user might never click on your site, but the value exchange still happens. You have no idea who is interacting with your website and you can’t access data to shape your strategy.

The challenge is this: after years of assigning budgets based on traffic and links, you’re faced with a measurement gap that threatens the value of search.

In other words, your content is feeding AI but you’re getting no transparency or benefit in return. Until new payment gateways for LLM crawlers come in, it’s only going to get worse.

A 2025 research report by TechnoLlama found ChatGPT alone has 800 million Weekly Active Users (WAU). Also, a 2025 Adobe Analytics report found traffic referred from AI to retail sites viewed 12% more pages per visit and had a 23% lower bounce rate. It’s no surprise OpenAI has rolled out shopping in ChatGPT amid rumours of a big Shopify partnership.

Having visibility on who is accessing your content using AI allows you to align your content strategy with this data and monitor the business impact.

That’s why we’ve created a simple Cloudflare AI Reporting Methodology that retrieves live data.

The key insights from this are:

Generative Engine Optimisation (GEO) traffic: You can see how many times your brand and content are exposed to a user when the AI bots visit your site for their purpose. This can be as granular as what is happening every second. This helps to provide you with the GEO equivalent to common SEO metrics such as impressions, traffic, and search visibility.

Traffic by AI crawlers

Popular content: You can identify the most popular content being consumed on your website. This means you can find other ideas that work in tandem with this content.

Most popular content consumed by AI

This article will walk you through:

  • How AI search works, and what you can track
  • How the Cloudflare AI Reporting methodology works
  • What business benefits our methodology delivers
  • How we think AI will evolve in the next few years

Read on to find out how you can leverage AI search before your competitors do.

How AI search works, and what you can track

How does Live AI search actually work?

Let’s see what happens when we put this prompt into ChatGPT:

What is the trendiest place to go in Mallorca this year for the summer? I want good vibes and go search the news to tell me.

Because we’ve asked ChatGPT for information about summer 2025, it’s retrieved recent information from the last few days. The top result is an article from The Sun from just four days ago (at the time of writing).

Screenshot from ChatGPT output

With recent information, ChatGPT enters the website live to learn the data in real-time. In this example, it looked at Bing’s APIs and grabbed the information from the relevant URL to learn more about the topic so it can source and reference it.

The advantage is it’s possible to track this process through ChatGPT’s agents – Open AI SearchBot, Chat-GPT User, and GPTBot. Then, once we understand which bot is crawling your website, we can see how many times your page is being viewed by individuals as they make queries on their favourite AI chatbot.

OpenAI user agents

An effective strategy for AI search is dependent on being able to see these patterns. Building content that connects to search intent has always required an understanding of your visibility in search and technical SEO. It’s no different to how traditional search has been measured and optimised.

How the Cloudflare AI Reporting Methodology Works

Our Cloudflare Methodology is currently one of the simplest ways you can see the consumption of your content in AI platforms.

Like many solutions, it started out as a customer problem. Our client wanted to find a way to measure the support ticket data AI gives users from their site in real-time. I decided to investigate whether ChatGPT leaves traceable tokens on the sites it crawls and whether this could be tracked using an existing tool.

Because traditional SEO tools like Ahefs and Semrush are unable to access live individual server logs without complex work to retrieve this data, I needed to find an alternative approach.

In my experiment, I used Cloudflare’s AI Audit feature to pull our site’s server logs to see if I could track AI bot activity through logs.

The process worked and I quickly refined it to provide the core insights businesses need.

AI User Visitors

Three use cases to generate immediate business value

The power of this data is using it to build an impactful content strategy. But how can you do that?

The answer is through combining LLM data analysis with commercial goals.

Let’s take a few common use cases that are instantly actionable.

1. Support deflection

The strategy is to reduce the volume of customer support tickets by encouraging customers to find answers themselves.

You can track ChatGPT hits on original documents to create a benchmark for experimentation. Then, you can refresh the self-service support documents and optimise them for LLMs. This will generate a lift in hits on the new documents and reduce support tickets on the same topics.

2. Brand lift

Brand lift provides insights into how people feel about your brand or product by enabling you to measure overall engagement due to AI search. If you are running a large brand activity, live information lets you see a spike in brand awareness correlating with your campaigns or other brand searches that need action.

Let’s look at an example.

When Sex and the City was released as a movie in 2008, the fact “Mr Big” had a heart attack on a Peloton was immediately reflected in search. Specifically, searches for “cancel Peloton subscription” increased by 809% and searches for “peloton safety” went up 733%.

With our Cloudflare methodology, Peloton would be able to see this change in AI search and serve the right content to move the user away from cancellation.

3. Editorial targeting

We can use data to see if the content we’re writing answers the questions that users are asking.

For example, we could find ways to repurpose our most popular post across channels to optimise reach. This may include commissioning a checklist, webinar, or even turning it into a virtual workshop.

Where our methodology stands out

Cloudflare AI Reporting methodology is a simple way to see how AI search is crawling your site.

It enables you to:

  • Connect insights to use cases that drive immediate short-term business value
  • Compare rankings on different LLMs like Perplexity and ChatGPT over time
  • See how many times your brand and content are exposed to a user with deep granularity, by minute or second
  • Identify the most popular content being viewed on your website
  • Understand which bot is crawling your website
  • Deliver data-backed validation for technical and content plans.

This is still a beta feature and since this article was originally written it has changed over and over again with more features.

Planning for the future

AI search visibility means you can generate quality content for your business.

You still have to ask yourself how you can continue to adapt to a rapidly changing environment in the longer term.

Businesses who are willing to change quickly – those who really understand AI and how their users search – will learn how to optimise for business outcomes.

The challenge right now is that AI search engines are moving from rapid response to a thinking approach. They will prefer faster websites that are more accessible from a technical perspective. While traditional SEO demanded a focus on technical issues 35% of the time, the rise of AI has increased this to around 50%. That’s because the goalposts have moved in terms of the external factors determining whether your site is crawled by agents.

Even more important, being able to track consumption of your content determines how you create a strategy to drive outcomes. Crucially, businesses not willing to integrate AI search data into their systems risk being left behind and this risk is rising at an incredibly rapid rate.

Get Your Free Cloudflare AI Reporting Playbook

Are you ready to see which content on your site is most popular with ChatGPT, Claude, and Perplexity? Get the step-by-step methodology right now.