How to Position Yourself on LLMs and Attract More Traffic in 5 Steps

Are you looking to position yourself on Chat GPT, Perplexity, or Gemini?

To do this, the first is to understand how visible your brand is on these LLMs (Large Language Models - the common name for Chat GPT).

Then, you need to implement various changes to your site to position yourself effectively.

This is exactly the method we will share step by step in this article.

For a few months now, we have been testing all the methods directly with our tool Qwairy.

Let's get started!

Step 1: Understanding what users are searching on LLMs

The first step to rank your brand in an LLM is to understand how users interact with LLMs like ChatGPT, Perplexity, or Claude.

This will help you understand the types of queries that will be entered into these tools. That's what we will teach you in this first part.

How do users interact with LLMs?

Let me give you a recent example.

During the Christmas holidays, my little brother received some fishing equipment.

He has quite a few questions about this equipment, and by chance, I was behind him and saw him opening Chat GPT.

His first instinct: he types in the fishing technique on Chat GPT.

Then he starts asking additional questions.

Unlike what one might do on a search engine, he is having a real conversation with the LLM.

But that's not the end of it.

He will then supplement the information he found by discussing with ChatGPT and doing a Google search.

To successfully position yourself in these LLMs, you need to develop a method to understand all the searches and queries that your potential prospects might be seeking on an LLM.

Unlike a search engine where you would just type in a query, we tend to have a genuine conversation with an LLM. And when we talk about conversation, it also means that we will use it to help us take a decision.

Typically, one might ask an LLM:

• What is the best fishing rod for pike fishing?
• How to choose your first fishing rod?
• Which fishing rod brand to choose for sea fishing?

To find out what questions users are asking LLMs, I will teach you different methods in this article.

What type of keywords are entered on LLMs?

All of them!

However we notice that many users tend to use LLMs for comparisons, often typing comparative queries.

Here are some examples of comparative queries that you can easily apply to your site:

• What is the best X for?
• Brand X vs Brand Y

Tip #1: Use Qwairy to Discover These Questions

I am the founder of Qwairy, a tool that allows brands to find the queries users enter on LLMs like Chat GPT.

The way it works is simple to explain: the tool understands your brand and identifies questions that users ask on an LLM.

It allows you to know all the comparative and transactional queries searched on an LLM regardless of your brand.

For example, let's take the brand Decathlon:

You can then track your visibility on LLMs week after week.

You can also manually add questions:

And to find these questions, I will give you three more tips that you can use for free.

Tip #2: Use Google Search Console with the REGEX Trick

Google Search Console is a goldmine for finding all the questions users entered into a search engine before seeing your site.

To do this, you need to use a Regex, which is as follows and shown in the image.

Here's the Regex to add:

^(who|what|where|or|how|that|which|when|why|how much|how many)\s

And here is the result on a website about craft beers:

As users tend to type conversational keywords and questions on LLMs, this type of question is worth following on a tool like Qwairy.

The only drawback is that these questions must have been entered on Google first.

Another Regex to add: ^(best|better|vs|versus|or)\s

→ This will give you all the queries containing the word "best."

Tip #3: Monitor Bing Webmaster Tools

Thanks to a partnership between Microsoft and OpenAI, we know that the results of ChatGPT, particularly its SearchGPT feature, are primarily based on Bing.

This is not done in real time. It seems that the tool relies on data from a few weeks ago.

The downside of Bing is that it does not allow Regex filtering, so you will need to manage that manually!

In the Search Performance tab, you will find information on the queries for which your site received impressions:

Tip #4: Questions from Perplexity

Finally, one last option that works very well is to rely on the other questions asked by the Perplexity tool.

Perplexity is a search engine that uses artificial intelligence to enhance its data.

It analyzes the Google results page to provide you with answers.

The great bonus of Perplexity is that when you enter a query into this search engine, it will give you similar questions surrounding that query.

And these questions are exactly the type of questions that people might enter into an LLM.

Here is what you might find:

Now that you know how to find the queries that people enter into LLMs, it's time to optimize your content to rank for those queries.

Several criteria will influence LLMs, including:

• the quality of your content
• the way you write
• Your authority on the web

By the way, in Step 3, we will show you how to discover the sources that LLMs rely on.

Step 2: Optimize Your Content

Once you've identified an important query for your business that users might enter into an LLM, you face two options.

Either you have no content that addresses this question, and in that case, the opportunity is to create content to rank for it.

Or you already have content that is invisible, and it's your competitors who are visible, and in that case, you will need to make some optimizations.

We explain what to do in both cases.

Case #1: You don't have a content piece for the Searched Query

If you don't have content yet on a topic or question that your prospect might be entering into an LLM, the process is fairly simple and closely similar to the classic SEO process.

The first thing to do is to create optimized content for that query or question.

Let's consider an example from a pool e-commerce website.

Prospects are likely searching for this kind of information on LLMs according to Qwairy:

• What is the best pool vacuum cleaner?
• What is the best heater for a pool?
• What is the ideal pH for my pool?

If you don't have any content yet, the first step is to create that content.

And to do this, it needs to be created in line with optimized SEO content traditions:

• A precise answer to the search intent that genuinely helps your users.
• Semantical optimization with a tool like Surfer SEO, YourTextGuru or Thot SEO (semantic importance is at least as significant on LLMs).
• Schema markup in your article, such as including an FAQ
• Information gain to outperform your competitors

If you're the only one addressing a subject and answering a question precisely, you will quickly position yourself within the LLMs.

Case #2: You Have Content But It's Invisible on LLMs

💡Note: There are currently two ways to be mentioned on LLMs. Either you're mentioned as a source, and your website links appears. Or you are mentioned directly as a brand.

This is a new paradigm shifting from classic SEO.

The first thing to analyze is whether your content provides a dedicated section specific to the query you wish to rank for on AI.

For example, if you want to rank for the query:

• What is the best pool vacuum cleaner?

It makes sense in your article to have a part that addresses it simply and effectively based on the most important criteria for choosing a pool vacuum.

If you don't appear, here's what you can do:

1. Analyze the Mentioned Source and reverse Engineer

If you do SEO, this is very similar to what you've been accustomed to this for years. It's a bit like looking at the SERP to check how to rank for a query and wondering how you can outperform others.

In this case, it's the same; we'll look at the source used by the LLM and analyse:

• What's the structure of the article?
• What sections are mentioned?
• What brands are quoted?
• How comprehensive is the article?

From there, you will optimize your content.

If you want to save time to discover the sources used by LLMs, you can find them directly on Qwairy:

In fact, they are even detailed with more granularity which can give you ideas for your backlink strategy:

2. Shorten Your Sentences

AI operates via tokens. A token generally represents a few characters (around 4). Logically, AI tends to favor texts that are simple to read and understand.

Initial tests indicate that writing in the passive voice and using straightforward sentences increases your chances of ranking in LLM's.

Also, note: Chat GPT uses a maximum of 4000 tokens per article. The content you place at the beginning of the article is therefore very important.

3. Analyze This Content on Bing and See How to Change It

The third trick that works very well is looking at Bing's results for your query. And do better than those who already rank there.

For this, it's simply a matter of using Bing's SERP!

Here, the method does not diverge from traditional SEO. You will simply look at the top results and figure out how to improve upon them.

4. Optimize Your Semantics

Optimizing semantic is just as essential for ranking on AIs as it is for ranking on search engines.

The principle of semantics is pretty simple. When you discuss a topic, it is natural to use certain terms that are close or related to that subject.

For instance, when talking about the French football team, it's normal to mention players like Kylian Mbappé. Search engines and AIs have developed algorithms that allow them to recognize these terms to gauge how coherent your texts are.

The advantage is that tools like Thot SEO or Yourtextguru or SERPmantics directly provide all related terms to help you position your content.

Aim for superior optimization and especially outperform your competitors on Bing! (Remember that Chat GPT uses Bing results!)

5. Add Schema Markup

Schema markup is a piece of code that you add to a web page to help search engines understand the content of that page.

What you need to understand is that for a bot, whether a search engine bot or an LLM bot, it's much easier to understand code than text.

For this reason, search engines created a language called schema markup. There are many different schema markups, but there's one that, in my opinion, is particularly interesting for ranking on LLMs.

This schema is the FAQ schema markup, which allows you to include an FAQ with schema markup in every article.

If you use a CMS like WordPress, it's straightforward to implement with the FAQ Schema plugin.

Another possibility if you have a different CMS: use this FAQ generator and add it directly into your HTML or by providing the code to your development team.

This enriches your article and makes the information understandable to search engines and LLMs.

However, sometimes all these techniques may not suffice. You may then need to become an authoritative source on the subject, and that's what we're going to discuss in detail in Step 3.

Step 3: Become an Authority Source

To explain this topic, we need to start with how AIs work.

AI performs a calculation: when this term appears, what is the term most likely to follow?

For example, if I say "a website to buy a pool vacuum cleaner is 'Site Name'."

The AI will try to calculate what is most likely to reappear after this phrase. And for that, it will rely on already available brand mentions in its database.

Thus, it has become extremely important to be a prominent brand to be cited on LLMs.

The Importance of Brand Mentions for LLMs

To position yourself on LLMs, it's crucial to have brand mentions across various sites. The goal now is no longer just to achieve a backlink but to have a brand that is mentioned.

Here's an example: for my blog about beerology, I wanted to position it on LLMs for the following question:

What is the best blog about beer?

This is, therefore, a question I added onto Qwairy (in this example it was in french):

And I can check the responses and see notably that I am mentioned on Search GPT:

I can discover the sources that AI uses.

Among these sources is the Bordeauxwineweb site where I obtained a link a few weeks ago:

It's a cheap link (5 euros) that I bought as a guest post, but it helped me get mentioned on this query:

With LLMs, backlink strategies are evolving.

The goal is no longer just to have a link on optimized anchor text but rather to have articles that will talk about you and include a link in the process.

With LLMs, we will ask for an article on the topic of the Best SEO Blog and rather request our brand to be included in it.

The more you increase your presence on other sites that support your brand concerning your activities, the better you will position yourself.

This represents a completely new paradigm emerging in link-building strategies.

Finding Sources with Qwairy to Obtain Brand Mentions

Another option is that, thanks to Qwairy, you can discover the sources for each question asked to the LLMs.

We can observe this, for example, with this query: What is the best beer expert training class where my brand is not mentioned.

As I have the sources, I can directly request those sources to include me in their accounts.

Or, I could create content on other sites to replace the currently existing sources.

Step 4: Optimize Your Technical SEO

Just like with search engines, it costs LLMs to visit and analyze your site. Therefore, you must work on your technical SEO and do everything you can to facilitate your site's crawling.

In this part, we will provide you with some simple improvements you can implement.

1. Implement Indexnow from Bing Webmaster Tools

The first improvement is very simple to implement.

Go to your Bing Webmaster Tools account, which is the equivalent of Google Search Console for Bing, and set up Indexnow.

Indexnow allows you to publish your content and have it indexed directly or more easily on Bing.

This is a straightforward improvement that will be important in the future.

2. Improve Crawling by Speeding Up Your Site

Option number two is to facilitate crawling by speeding up your site.

Certainly, if your site is slow, LLMs and search engines cannot spend as much time on it as they can on other sites.

This is especially critical if you are a site with substantial volume and thousands of pages. To speed up your site, the first thing to check is whether your server is adequately sized to handle traffic spikes. Avoid shared hosting for a few euros and look for something a bit pricier.

The second thing, which is quite simple to implement, is to convert all your images to WebP format and reduce the size of these images to around 100 KB.

You can then implement additional measures, such as minifying JavaScript or optimizing CSS with your development team.

3. Redirect all 404

Another very easy technical improvement to implement is to eliminate 404s. At times, throughout a site's history, we may add content and later decide to remove it.

It can also occur that pages are automatically generated as 404 for some reason.

The problem is that these removed contents must absolutely be redirected to new pages to prevent LLMs and search engines from wasting crawling budget visiting useless pages.

To find the 404s, a very straightforward method exists in Google Search Console:

Additionally, you can turn to a crawl using a tool like Screaming Frog or Website Auditor.

And above all, log analysis allows you to complete this assessment!

4. Analyze Your Logs to See When and Where LLMs Visit

To take it a step further, the rise of LLMs is an opportunity to implement log analysis.

A log analysis entails examining the log files generated by servers, applications, websites, or IT systems.

This will allow you to know when LLMs visited your site and which pages they accessed. You may be surprised to find they frequently visit pages that are of no use to you.

One potential improvement is to add these pages to the robots.txt file and indicate to the bot not to visit them.

5. Introducing the LLM.txt File

The llms.txt file is a standardized Markdown file placed at the root of a website (for example, https://yoursite.com/llms.txt). It serves as a guide for large language models (LLMs) by providing a concise and structured overview of the site's content.

It's similar to the robots.txt file but intended for LLMs.

Currently, we have no certainty that LLMs uses this file, as it remains primarily a proposal from various SEO consultants to create such a file. However, it's possible that it may be used in the future.

Step 5: Implement Traffic Tracking from LLMs

Monitor Your Visibility on LLMs with Qwairy

Qwairy.co also lets you keep tabs on how visible you are across large language models.

Simply set up the queries you want to monitor, and you'll get weekly updates on how you're performing.

You'll also get details by providers performance (Gemini, GPT, Sonar, SearchGPT…) and by keyword:

You can even take things further by comparing your performance with that of your competitors:

Track Traffic on Your Site from LLMs

To assist our clients, we will set up tracking of traffic stemming from LLMs. We will directly create a Looker Studio dashboard that will be accessible to all.

You can, if you wish, utilize this, and here's what it looks like:

With this management, you can observe:

• The individuals who arrived at your site from LLMs and from which specific LLM
• The landing pages on which people entered
• Day-to-day traffic from LLMs

You might find that a certain page is quite popular on LLMs:

The dashboard is connected to Google Analytics data.

To utilize it, simply make a copy of the dashboard directly.

Click here to access the dashboard

Remember to make your own copy like this: 😉