7% of B2B buyers reported using LLMs like ChatGPT in their buying process in 2025, and it’s safe to assume that number will only keep on growing.
On top of that, the referral traffic ChatGPT sends to websites often converts better than Google traffic. According to Kevin Indig:
“As of June 2025:
- ChatGPT’s conversion rate of transactional traffic was 6.9% in the US compared to 5.4% for Google.
- In the UK, ChatGPT reached 5.5%, which is on par with Google.”
ChatGPT sends higher quality traffic to websites, at least in the US” – with “higher quality” meaning higher intent.
This makes sense:
Users self-educate by “conversing” with ChatGPT, and when they click out, they’ve found their answer and are ready to buy. If you show up in ChatGPT answers, you’re likely to get on your prospects’ shortlist, and if you’re lucky, you might already be their #1 choice the moment they click through to your site.
Ranking in the Search results used to be enough for many brands. Today, your content also needs to appear in Google’s AI Overviews and in LLM-driven search tools.
ChatGPT: the Basics
Before we dive into how the rise of ChatGPT has influenced Search, and how you can influence ChatGPT, it’s important to understand the tool’s place in today’s Search landscape. For anyone operating in the Western world, that means looking at two things:
- How ChatGPT compares to other popular LLM tools.
- How answer generation by ChatGPT differs from result generation by Google.
ChatGPT vs other popular LLM tools
The key points in which ChatGPT differs from most other popular LLM-driven tools are that it:
- relies on OpenAI’s own LLMs (GPT-4o and GPT-5).
- pulls its Search data from Bing and, most likely, a combination of OpenAI’s own index. Potentially uses Google Search results, and other third-party APIs.
- supports real-time browsing.
It has a lot of these characteristics in common with Copilot as Copilot runs on OpenAI’s GPT-4 models. However, there is no cut-off date for Copilot’s training data, whereas the training data for ChatGPT’s most recent LLM (GPT-5) cuts off at June 2024.
Battle of the browsers
Already an alternative to Google Search, LLM companies are also going after a piece of the search infrastructure pie. Perplexity launched its Comet browser in July 2025 and OpenAI followed with Atlas in October of the same year.
In the meantime, Google is rolling out Gemini in Chrome and incentivizing Chrome users to click through from AI Overviews to the conversational AI Mode interface – both powered by Gemini.
Toppling Chrome as a browser market leader will be hard, but if anything, the last few years in Search have taught us that to stay ahead, we need to stay on top of the latest evolutions.
External influences
The companies behind these LLM tools don’t exist in a vacuum. Some of them receive external, others internal funding, and all have partnerships with publishers.
Through publisher partnerships, OpenAI gets direct access to the publisher’s archives and, sometimes, fresh and even paywalled news in exchange for payment, citations, and often also other deals (access to OpenAI technology, advertising deals, …).
One could think that ChatGPT would prioritize citing content from partners in its answers, but it’s not as simple as that. Even if it wanted to, the tool still relies on the sources it gets its data from, and that includes Google.
When Google removed the num=100 parameter from its search around September 10, ChatGPT couldn’t just scrape Reddit pages ranking lower than Page 1 anymore, and Reddit citations dropped significantly.
Answer Generation by ChatGPT vs Google
ChatGPT doesn’t just provide answers in a different way than Google Search does, it also bases those answers on different data. Whereas Google uses its own search data, ChatGPT relies on a mix of publicly available data, licensed data, and data created by human reviewers as training data for its LLM, and ChatGPT with Search relies on that training data plus search data coming from Bing.
No ranking mechanism
The number one difference between how ChatGPT and Google provide answers is that Google has an explicit ranking mechanism and ChatGPT does not. The latter generates answers based on your specific prompt, its training, and statistical patterns.
Like other LLM-driven search tools, ChatGPT looks for the most relevant (or predictable) content.
The Influence of ChatGPT on Search
When ChatGPT was launched in late 2022, it introduced a new way for people to discover content and gather information before making a purchase. Instead of having to click through the Search results, decision makers and influencers can now simply tell ChatGPT what they were looking for and receive an answer based on its analysis of multiple sources.
ChatGPT can speed up the research process significantly, especially for those needing to check a lot of boxes before they can shortlist a product or service provider.
And buyers have taken note. Between the summer of 2024 and the summer of 2025, ChatGPT visits grew from 3.5 to 6.8 billion in the US (+94%) and from 868 million to 2 billion (131%) in the UK (Source).
And these visits translate to referral traffic, too. In the same period, ChatGPT’s referral traffic to websites grew by 3,496% for the US and by 5,950% for the UK (Source) until it dipped significantly between July and August 2025 (Source). It remains to be seen how this will evolve as OpenAI experiments with the sources it pulls from and presents.
Given its strong growth both in visits and outgoing traffic, it’s no surprise that an analysis of 3,000 websites by Ahrefs in January 2025 showed that of the websites receiving traffic from AI (63%), ChatGPT was the biggest referrer, accounting for 50% of AI traffic.
So while you may still be seeing overall low traffic from ChatGPT at the time of writing, the tool is becoming an increasingly important touchpoint in the buyer’s journey.
Does this mean Search is dead? Quite the opposite. The introduction of ChatGPT and other LLM-driven search tools has meant a diversification of Search and the practice of SEO.
It is no longer enough to show up in the Search results. Today, your brand also needs to show up in AI answers.
Before we go into strategies to increase your brand’s visibility in ChatGPT, it’s important to be aware of how ChatGPT differs from other popular LLM tools, and what the differences are in how it generates answers in comparison to Google Search.
What is GPT SEO?
GPT SEO or ChatGPT SEO refers to the strategies and techniques applied to optimize content in such a way that it is more likely to be included in ChatGPT answers.
How to do SEO for ChatGPT
Despite the differences in answer generation, optimizing your content for ChatGPT involves applying a lot of the same techniques as those used in modern search engine optimization. At the same time, ChatGPT optimization requires a more holistic marketing approach.
Let’s have a look.
1. Build an omnipresent, consistent brand
Use the same phrasing to describe your brand and your products on your website, your social media profiles, and on review sites such as Capterra and G2.
Formulate brand and product descriptions exactly how you would want ChatGPT to repeat them, which should be the way you want to be known by your target audience. To maintain consistent messaging as you expand your presence across new channels, you can create an internal document with the most-used brand phrases.
Aside from aligning how your brand appears across web properties, it’s also important to expand your digital footprint through media appearances, being active on relevant forums, sharing user-generated content, and creating new-for-you types of content.
2. Build quality backlinks
A study by Kevin Indig and Semrush found that backlinks influence a site’s visibility in AI-generated answers. More specifically, they learned that:
- the quality of the links matters more than having lots of backlinks.
- the number of domains linking back to a site is more important than its total number of backlinks.
- nofollow are just as valuable as follow links.
- image links are at least as powerful, if not more powerful, than text links.
Aside from implementing some of the more traditional link-building techniques, B2B brands can use tools such as Ziptie.dev and Peec.ai to find web pages that ChatGPT cites as sources for answers related to their product to then try to get included in those sources. Some ways to do that are:
- by offering the brand or writer a quote or insight.
- by paying for a sponsored placement. This is especially the case for review sites and high-authority domains.
- by suggesting a link trade.
Alternatively, you can check what type of content is cited and use that as inspiration to create your own post or page that can be included as a source.
3. Optimize for your audience’s questions first
According to Helene, the most important thing to take into account when optimizing for ChatGPT is that, at the end of the day, you’re still optimizing for humans. ChatGPT isn’t the one making a purchase through your website. It’s a person who’ll do that.
She recommends creating unique and hyper-specific content that targets your audience’s pain points, questions, and doubts. This type of content establishes trust and authority, and because the tool appears to favor unique and original data, it is more likely to be picked up by ChatGPT.
Another reason to be more specific when optimizing content for ChatGPT is that users search differently there than they do on Google. Kevin Indig found that the average length of a prompt used for ChatGPT is a lot longer – a lot more specific – than the queries used for Searches on Google.
4. Pay attention to content structure and formatting
It’s easier for ChatGPT to interpret your content correctly when it’s well-structured. That includes:
- following the correct heading hierarchy.
- using bullet points to add semantic interpretation.
- including tables to visually represent information.
5. Implement technical SEO
ChatGPT needs to be able to access, read, and process your content to be able to include it in its answers. Luckily, the best practices in this case are the same for ChatGPT SEO as for traditional SEO:
- Check that your Robots.txt and firewall settings (for example, through Cloudflare) aren’t blocking any search engine or AI crawlers.
- Make sure that nothing that needs to be indexed is hidden behind a paywall or a login.
- Clean up duplicate content.
- Fix 404s and broken redirects.
- Tidy up redirect chains.
- Be smart with how you set up internal links to let search engines and LLMs know what your site is about and what you are an authority in.
Be careful with schema
Adding schema is an important part of technical SEO. Unfortunately, it’s easy to get it wrong. On top of that, ChatGPT tokenizes content when it processes it, which means that whatever schema markup you added is destroyed again. (Source 1, Source 2)
In other words: for now, it doesn’t seem like adding structured data to your content has a positive effect on ChatGPT visibility.
If you’re certain you can add it correctly, go ahead as it can improve your organic search rankings. But if not, it’s better to not add schema than to risk doing something that might negatively impact your site.
6. Improve your rankings in Bing
As ChatGPT relies on Bing for its data, ranking well on that search engine may increase your chances of being mentioned in ChatGPT.
7. Experiment, don’t chase
We’re still in the early days of ChatGPT SEO. There is so much we don’t know yet, and what works is changing rapidly as OpenAI experiments with the sources it pulls from.
If you have the resources, it’s good to try things out, but be careful not to get caught up running after every new trend trying to hit quick wins.
| ChatGPT is only one part of the shifting AI search landscape. If you’re building a broader strategy, compare how dynamics differ in SEO for Perplexity, how Bing-connected discovery affects inclusion in SEO for Microsoft Copilot, and how Google’s ecosystem changes the game through AI Overviews and AI Mode in SEO for Gemini. |
Tracking ChatGPT Search Performance
As Helene explains, attribution is still very tricky as we don’t have a search console for ChatGPT. You are only able to view users who click from ChatGPT directly to your site (when it’s cited as a source). Even then, GA4 may categorize that click as organic or direct traffic depending on the user’s behavior:
To get an idea of the traffic ChatGPT sends you, you can set up a GA4 report in the following way:
1. Dimensions: >page referrer
2. Metrics > sessions/conversions/engagement rate/… – Choose this to your liking
3. Filter > regex contains:
| ^https:\/\/(www\.meta\.ai|www\.perplexity\.ai|chat\.openai\.com|claude\.ai|chat\.mistral\.ai|gemini\.google\.com|bard\.google\.com|chatgpt\.com|copilot\.microsoft\.com|edgeservices\.bing\.com)(\/.*)?$ |
Easier to set up and visually clearer to use is our free LLM dashboard. It shows you referral traffic from LLMs like ChatGPT and Perplexity, which pages this traffic lands on, and how it converts.
Some popular SEO tools like Ahrefs now list how many AI citations a web page has, but we found their data to be lacking.
At Flow, we use the lesser-known tools ZipTie and Peec, but also with those, you need to take into account that they simulate data and run queries for averages. Tracking actual referral traffic is still the best way to know how many people are visiting your website through ChatGPT.
Don’t Stay Behind
B2B buyers increasingly turn to ChatGPT as part of their research and decision-making process. Marketers can seize this shift in the Search landscape as an opportunity to gain visibility and build trust throughout the buyer’s journey by investing in SEO for ChatGPT.
On top of implementing traditional technical SEO best practices, it’s more important than ever to build an omnipresent brand through consistent messaging.
Get in touch to discuss how our AI visibility audit helps you define areas of improvement and how our SEO strategy services integrate ChatGPT optimization into your overall marketing strategy.
