Ranking in AI Search—Practical Guide

If you find this article in any normal search engine, such as Google or Bing, I suggest you close this tab and search “How to Rank in AI” again.

But if you’ve discovered this article through Perplexity, GPTs, or any other AI-powered search engine, congratulations! You’ve found the masterpiece.

Before we discuss this new method of ranking in AI, I need you to forget all those old school website SEO clichés. Because as a sailor, you wouldn’t find a new continent with an old star map.And a deeper thing is that most people don’t know the AI search engine is fundamentally different from modern search engines like Google or Bing. They are not just databases with web scrapers; they are scrapers with minds. I will explain the difference at every step and tell you why you should do this to rank in AI (when I say AI, I usually refer to those large language models trained based on neural networks).

Also, this is an experiment to test my hypothesis. I will use this article to prove it and even add more paragraphs in the near future, because after talking to these LLMs every day, I gradually believe that ranking in AI will be a very important issue in the future, just like how SEO has become common in today’s world.

So, buckle up! I’m about to start with my first theory.

Multimodal strategy

Unlike search engines that primarily rely on text-based content and backlinks, modern and postmodern models are capable of processing and understanding information across various modalities: text, images, audio, and even video.

This is because they are built this way, and multimodal inputs help large language models reason better.

AI models are trained on diverse datasets, allowing them to understand context and nuances across different media types. By providing information in multiple formats, you’re giving the AI more data points to work with, increasing the likelihood of your content being deemed relevant and informative.

Implementing a Multimodal Strategy

One Topic, Multiple Media

Begin by crafting clear, informative text that forms the foundation of your content. Then, visualize key concepts using infographics, diagrams, and relevant images that reinforce and expand upon your written information. Don’t stop at visual elements; consider creating audio supplements like podcast-style discussions or concise summaries of your main points, catering to those who prefer auditory learning or consume content on-the-go.

Short, informative videos can offer quick overviews or demonstrations, providing yet another way to engage with your topic. If your content includes data or statistics, present this information in both textual and visual formats using graphs and charts, making complex information more accessible and understandable. By consistently presenting your topic across these different modalities, you create a comprehensive, multi-faceted resource that not only caters to various learning preferences but also provides AI search engines with rich, diverse content to analyze and rank.

If you feel lost, don’t worry. I will give some examples that are also applied to this article.

• Text: I created this article with a well-structured format.
• Image: All the images I used are well-formed with alt descriptions, and these descriptions are also contextual. Another method is to create a summary thumbnail for this article.
• Audio: For the current topic, create a related audio podcast or a simple voice introduction. You can embed it within the blog post or put it in a podcast. Here, I recommend NotebookLM. Below is the podcast for this article.
• Video: Create a video about the current topic.
• PPTX/PDF: This is not a common approach, but if you frequently use AI search engines, you’ll find that they also index PPT and PDF content shared online. Of course, if you can publish a paper using company resources, that would be even better.

After you have completed these content materials, we will move on to the next stage – distribution.

One brand, multiple touchpoints

This is different from the old SEO strategy. You may have been taught that your domain website is the most important. That’s why there were so many articles about how to add internal links and make sure the CTR, etc., were optimized.

However, in the reality of AI search, everything here has changed. When you search for something like “best coffee bean for mocha,” you only get answers, not specific websites. So we can conclude that the boundaries between URLs and domain names have become blurred under large language model-based searches.

This is why distributing your content across multiple platforms is crucial for maximizing visibility and engagement.

• Increased Data Points: AI models crawl and index content from various sources. By spreading your brand across platforms, you create multiple touchpoints for AI to recognize and associate with your content.
• Diverse Context: Different platforms provide unique contexts for your content, helping AI understand its relevance in various scenarios.
• Cross-Platform Validation: Consistent presence across platforms can signal to AI that your content is credible and widely accepted.

The following are the steps to take:

1. Identify Relevant Platforms

• Social Media: LinkedIn, Twitter, Facebook, Instagram
• Professional Networks: GitHub (for tech), ResearchGate (for academia)
• Content Platforms: Medium, Substack, YouTube
• Q&A Sites: Quora, Stack Exchange
• Podcasting Platforms: Spotify, Apple Podcasts, Google Podcasts
• Sharing Platforms: Google Docs
• AI Content Platforms: Perplexity Page

2. Tailor Content for Each Platform

• Adapt your core message to fit each platform’s format and audience.
• Use platform-specific features (e.g., hashtags on Twitter, long-form articles on LinkedIn).

3. Cross-Link and Reference

• Create a web of connections between your content across platforms.
• Use consistent branding and messaging to reinforce your identity.

This strategy increases the brand’s digital footprint and creates multiple touchpoints for AI search engines to discover and index the content.

Speak Their Languages

When optimizing for AI search, it’s crucial to consider two types of languages:

1. Human Languages: The diverse languages spoken by your target audience.
2. AI-Preferred Language: The language structures and formats that AI models are trained on and respond well to.

You may notice that when you input content in different languages or formats, you receive different answers from AI systems. This is due to variations in the training data and text processing methods used for different languages and cultures.

Therefore, your content should cover as many languages and cultural usage habits as possible. Fortunately, AI is very good at this, and you can use it to help you with batch translation and review.

Regarding AI Language Preferences, I currently don’t have any specific ideas or actual cases. However, I’ve noticed that a clear text structure and some Q&A-like descriptions might enhance their indexing of your content.

Get indexed

Yes, essentially, all the steps we take are to get AI to index our content. But this is where it gets very interesting. Most large language models claim that they regularly collect and index content from the internet, but very few disclose their specific channels and standard norms for indexing and acquiring information. However, what we do know is that your search records and behaviors are documented. These might even be used to fine-tune another model in the future. Here you’ll find that maintaining content update frequency and creating content earlier often becomes more important. Suppose the AI’s database is from April, and you only created content in May, then naturally your brand won’t be indexed by the AI.

Another method is what I consider theoretically feasible but have never practiced. I call it “flood indexing”. To some extent, you can create a large number of bots through automation or manipulate your users to constantly ask questions on AI search engines, establishing a strong connection between you and certain keywords.

Again, I wouldn’t try this. I would attempt to do something milder within the rules. For example, creating related public Perplexity pages and shareable AI question threads. Of course, this is all a black box, but isn’t Google the same?

Conclusion

The original intention of writing this article was to help SMBs gain greater exposure and opportunities. However, Google’s SEO has become increasingly complex and cumbersome, with each core algorithm update hurting some independent sites and businesses (if you search “why my traffic drop” on the internet, you’ll know what I’m talking about). After talking to many people, I also found that people’s search habits are indeed changing with the birth of LLMs. I believe this is a trend.

This trend is that in the future, people will use new LLM-based searches to obtain information, and there will also be LLM-based optimization (GEO) in the future. As for traditional SEO methods, I believe they are consistent with AI ranking to some extent, because large language models themselves are also algorithms, just more advanced algorithms than Google’s retrieval. But we can see some trends, such as the disappearance of the programming tutorial SEO industry, indicating that such a future may not be far from us. Regarding this content, it will be the first article on ranking in AI, but definitely not the last. I will continue to update and optimize until you can see me on AI search engines.

Q&A

1. Will AI search replace Google?

Yes, AI search is likely to replace traditional search engines like Google in the future. While Google is incorporating AI into its algorithms, pure AI-driven search offers more personalized, context-aware results. As language models advance, they’ll better understand user intent and provide more accurate, comprehensive answers. This shift is already visible in certain sectors, like programming tutorials. Preparing for this AI-dominated future in search is crucial for maintaining online visibility and relevance.

2. What is search intent?

Search intent refers to the purpose behind a user’s search query. It typically falls into four categories: informational (seeking knowledge), navigational (looking for a specific site), commercial (researching products/services), and transactional (ready to make a purchase). Understanding and aligning with search intent is crucial for both traditional SEO and AI search optimization, as it helps provide more relevant content to users.

3. Why is multimodal media important?

Multimodal media is crucial because it shapes AI’s understanding and output. For instance, Midjourney’s astronaut images often include the American flag, indicating future growth and marketing can be optimized through multimodal models. Future searches may not be text-based. If you’re a hiking boot brand, your logo might appear in AI-generated hiking videos or images when people use AI for content creation. This opens new avenues for brand visibility and engagement in an AI-driven digital landscape.