This is a deep dive in the inner workings of ChatGPT and Perplexity. The goal of the post is to answer a simple question:

Can you get your job board to rank on AI-powered search platforms like Perplexity or Chat-based LLMs like ChatGPT and Claude.

Why is it beneficial for your site to appear on ChatGPT and Pexplexity?

People have been reporting seeing traffic from Perplexity and ChatGPT as a source in their analytics. 

LinkedIn influencers have claimed for over 2 years now that ChatGPT and Perplexity will surpass Google in organic traffic and search volume. 

Although this is grossly exaggerated, it is not entirely wrong that something is changing. The more consumers adopt these tools, the more often they use them to get answers and provide references, hence skipping a Google search. 

Google has lost some market share in Search for the first time in years, dropping from 91.6% to 89.7% in 2024. This is not a dramatic change because most of it was driven by Bing’s publicity stunts with ChatGPT in 2024, but still, it is the start of something.   

It is also worth mentioning that these tools often answer without requiring users to click on a link. This hurts direct organic traffic but positively impacts your brand’s organic performance as people learn about your brand name. 

Google has also rolled out AI-generated previews for certain query types, which will further reduce organic traffic in the future. 

In summary, you generally want your site to appear on any LLM or AI-powered search engine. This will positively impact your organic traffic and might be just enough to compensate for Google's traffic loss. 

How to get traffic from ChatGPT and Perplexity

I am seeing agencies promising to rank your site in LLMs and Perplexity. Before you spend money on this, let’s do a deep dive and see how much you can influence the chance of your site appearing on these platforms.

Although my target group is mostly job boards, what is written in this post applies to any site. 

It is time for the more technical part of the article. We will start by explaining how ChatGPT generates search results, using this as a foundation for the rest of the article. We will then build on that to explain what Perplexity does differently, and then we can look into strategies for influencing this. 

How does ChatGPT work for search?

Traditional search engines (like Google or Bing) crawl and index billions of web pages, ranking them based on various signals, such as authority, backlinks, relevance, etc. This is usually known as the “Google” algorithm (I am ignoring Bing because no one in their right mind optimized for 3% of the traffic). 

Here is what I will get today if I search for “ruby on rails jobs” on Google: 

There is a mixture of both free (what is called organic) and paid (sponsored) search results. 

As we already know, LLMs work a bit differently. They’re “trained” on massive text datasets, which helps them generate conversational responses. 

So, if someone does a search on ChatGPT for “websites to find ruby jobs, the results will be significantly different:

Notice that I am explicitly not using ChatGPT's browse feature, so the answer will use the data from the LLM. 

In this case, the LLM generates text based on my input. Thus, ChatGPT's training data frequently mentions these job boards after queries like “looking for rails jobs.”   

There is a difference in the output:

Google returned Ruby on Remote as a first source (a remote Ruby-on-Rails) job board, while ChatGPT decided to go with "Rails Jobs."

How does Perplexity work for search?

We need a bit of history lessons here.

History of Perplexity

Perplexity distinguishes their platform from traditional search by referring to it as an “answer engine.” 

The guys at Perplexity initially worked on enterprise search, and the story is quite interesting. As the CEO outlines in this video, modern search engines are bad at delivering answers to his team's questions, which he searched for during their daily work.

In this video, he clearly states that ads and SEO heavily manipulate Google, so the top results are not helpful. 

The solution: let’s use an LLM. 

But, in his own words, he quickly learned that LLM outputs “cannot be trusted,” so they needed something to verify them. Whenever he received a response, he would consult others and discuss whether the response was accurate. 

Frustrated by these inaccurate, outdated answers, GPT-3 could return, his CTO connected GPT-3 to a RAG system that queried Microsoft Bing’s index, dramatically improving answer quality.

So yeah, the system (search index) initially discarded as heavily SEO-optimized and untrustworthy is now part of the ranking and validation system.  

The actual search mechanism of Perplexity

This is only based on my research. Honestly, there is no way to fully understand how Perplexity works, so use this information with caution. 

This technology combines two powerful tools: traditional search engines and AI language models. The search engine finds relevant, up-to-date information online, while the AI understands and processes language like a human would. They create comprehensive answers by combining the AI's ability to understand and communicate with the search engine's vast, current knowledge base.

When these tools have real-time access to the web, they still reference search results from a search engine index, but they also use advanced language understanding to pull together summarized answers from the search index, not just lists of links.

So, in summary, the results and the order of the search index still play a significant role in what is mentioned in the answer. 

How can you get your site referenced in LLMs and AI search

Now that we have covered the basics of how these two different systems generate search results, we can discuss how you can influence these. 

Let’s start with the LLMs.  

The more often your job board name appears in the LLM's training data after the targeted keywords, the higher the chance your brand will appear as a summary. 

We are after quantity here. If you are a very popular name, your job board will be mentioned many times about many keywords, so the chance of the LLM triggering the correct sequence after “how to find rails jobs …” that includes your brand is based on frequency. 

The more popular brands will have a higher probability of appearing in an LLM. The higher you score in the ranking, the higher your chance of appearing in an LLM. 

There have been a couple of studies (the one in the screenshot is the latest) that confirm this correlation:

“Brands ranking on page 1 of Google showed a strong correlation (~0.65) with LLM mentions. Bing rankings also mattered but less so (~0.5–0.6)”.

Last, you have to consider one more thing: LLMs are static. They have a knowledge cut-off, and if your brand was created before this date or you did not have many mentions, your chance of appearing is zero. You have to wait for the subsequent retraining. Sorry. 

Traditional SEO has a significant impact on this relationship.

Specifically, backlinks and digital PR campaigns targeting your brand keywords will significantly impact your ranking on search engines, as will all other traditional SEO measures.  

How can you impact your AI search ranking (Perplexity, ChatGPT with Browsing)

Perplexity is a strategic fusion: search engines provide real-time data collection and organization, while the LLMs deliver sophisticated language processing. 

Together, they create an optimized system for generating relevant, well-articulated responses.

Still, they need some indication of which links are qualitative and should be considered. Ranking position is a strong indicator for them. 

Hence, strong organic rankings correlate with better visibility in AI-driven chats since top-tier content will likely appear in both the LLM response and the organic search results. 

Strong organic rankings correlate with better visibility in AI-driven searches. Search rankings influence AI Searches significantly but are not the whole story.

However, it is not as simple as “if you rank in one, you rank in all.” LLM search and chat interfaces use different retrieval methods, rely on various training data sets, and display citations (if at all) in other ways.

LLMs also use vector-based similarity, so a page with excellent structured data or content that concisely answers a question could be highly favored by an AI-based retrieval system—even if it’s not #1 in organic search. 

Still, the page must be considered part of the initial search result list. 

So, despite these differences, ranking highly in organic search is still highly advantageous. You’re more likely to:

1. Be part of the data that LLMs ingest or reference (especially if they connect to real-time search APIs).
2. Be recognized as a “trusted” and relevant authority that might get pulled into AI-generated answers.

But that doesn’t directly translate into always being the primary or exclusive source in an LLM-generated response. 

Don’t forget – LLMs are still “wild” by nature, and every response has a certain amount of randomness.  

The bottom line is that, as of now, anyone claiming that they will influence your ranking on ChatGPT or Perplexity is not telling the whole picture:

Can agencies help you rank on Perplexity and ChatGPT (fast)

Well, this is quite simple.  

The short answer is NO.

If you are not already ranking in an LLM, the only way to change it is to build solid backlinks with your targeted keywords as anchors and wait for the subsequent retraining of the LLM. 

Still, even then, there is no guarantee because you don’t know how “much more often” your competitors are mentioned for these keywords. 

Regarding Perplexity and ChatGPT with browsing, getting a brand to rank in the queries is more or less correlated with traditional SEO. 

A high organic position will likely improve your chance of being mentioned for a specific query. 

Again, this does not happen overnight - SEO takes time, so the answer is no to fast and yes if it is a good, proper agency.

Future perspective – ranking on Perplexity and ChatGPT

The outlined strategies will work for now, but will they be sufficient for the future? Good question.

We already know that Perplexity and OpenAI scrape the internet aggressively (and usually ignore the robots.txt preferences). 

Although OpenAI's explanation is that they need new data to trail their LLMs, Anthropic's reasons might be slightly different.

One reason for scraping is related to user searches—whenever someone searches, Anthropic gets the data from these pages and uses it as input for the LLM. 

Another likely reason is that they are building their search index. I don’t think this is a secret anymore. 

According to Srinivas, Perplexity has strategically invested in search infrastructure independence, developing its comprehensive stack that includes custom crawlers, indexes, and a proprietary PageRank algorithm. He has repeatedly highlighted the effort of crawling and re-crawling websites to obtain the most up-to-date information. 

It would look a bit more like this:

This speaks to another optimization angle: formatting and structuring source texts to optimize the retrieval stage. In the world of SEO, we call this structured data and schemas. You already know the job posting schema, but there are over 30 different schema types. Start exploring them.

Therefore, there are many new unknowns. Perplexity might show aggregated or summarized data from multiple web pages, or it might “prefer” specific sources due to the nuances of its retrieval model and internal ranking. 

The convergence of diverse data sources with LLMs creates unprecedented information processing and knowledge synthesis capabilities. However, this integration also introduces new complex risks in information quality. 

Strategy Implications – Getting your Site to Rank on LLMs and AI Search

1. Traditional SEO impacts search engine ranking and is more critical than ever.
2. High-quality backlinks with proper a-href anchor text related to your target keywords are critical.
3. Structured Data and Clear Summaries: Marking up content (FAQ schema, how-to schema, etc.) helps both classical search engines and potential AI-based summarizers better parse your content.
4. Monitoring citation practices on Perplexity for your keywords and niche can also indicate how to optimize for AI search. 

Good SEO and authoritative content creation remain essential foundations for being “seen” by any search or retrieval system. However, expecting a direct 1:1 mapping—i.e., being #1 on Google guarantees top mention in an LLM-based chat—is oversimplified.

Instead, think of LLMs and AI chat tools as a new layer of discovery that overlaps with organic SEO but has its nuances, rules, and algorithms.

Are you seeking a specific consultation with me to discuss SEO optimization tactics for your job board or AI-based automation? Book a call!