AI optimization: how to get cited by ChatGPT and Perplexity

Why AI optimization is now a necessity, not an option

Just a few years ago, the phrase "AI optimization" sounded like a futuristic curiosity. Today, it is a daily reality. Gartner predicts that by 2028, traditional organic traffic from Google will drop by 50%, and zero-click searches already account for 60% of all searches (source: OtterlyAI, Generative Engine Optimization Guide, 2025). This means that more and more people, especially in B2B, are asking questions directly to ChatGPT, Perplexity, or Bing AI and expect ready-made answers, not a list of links.

If your company is not present in these answers, you are losing clients to competitors who have understood the mechanism. AI optimization (GEO: Generative Engine Optimization) is not a fad but a necessity if you want to remain visible in the digital landscape. In this article, I will show you how to prepare your content step by step so that it gets cited by language models.

Importantly, AI optimization does not require abandoning traditional SEO. On the contrary, good SEO practices such as website positioning are the foundation on which visibility in AI is built. The difference is that AI does not read pages like a human: it needs structure, clear data, and authority.

How generative AI engines work and what they cite

To effectively optimize for AI, you need to understand how language models work. LLMs (Large Language Models) do not search the internet in real time: they rely on training data that includes vast collections of texts, including articles, company websites, forums, and social media. When a user asks a question, the model generates an answer based on this data, often citing sources it deems credible.

It is therefore crucial that your content is: (1) present in the training data, (2) well structured semantically, (3) perceived as authoritative. According to OtterlyAI (2025), brands must be present in AI answers through citations, statistics, links, and mentions, not just in blue links. This means that having a website is not enough: you need to be where AI draws its knowledge from.

One of the most important elements is Schema.org: structured data that makes it easier for AI to understand and cite content. With it, crawlers can read that a given page contains an article, reviews, contact details, or FAQ. This increases the chance that AI will choose your content as the source for an answer.

It is also worth remembering that AI often uses UGC (User Generated Content) such as Reddit. Being present there increases the chances of being cited. Therefore, it is worth considering activity on industry forums, even if it does not bring direct traffic.

Practical steps for optimizing content for AI

1. Build topical authority

AI does not cite random pages. It needs sources that are credible. So the first step is to build authority in your niche. Write regularly, answer specific customer questions, publish research and data. According to a study by Vieira et al. (2019), owned media and inbound marketing have a greater impact on B2B customer acquisition than paid media, and the effect of investment grows over time: it becomes significant after 3-4 periods. This is a long-term game, but it brings lasting results.

In practice: instead of writing generic articles about “B2B marketing,” create detailed guides, case studies, and analyses. Answer the questions your clients ask in sales conversations. These are the types of content AI will cite.

2. Use structured data (Schema.org)

Technical GEO is the foundation. Add Schema.org markup to your site, such as Article, FAQPage, Organization. This helps AI better understand what your page is about and what information it contains. It increases the chances of being cited in answers to specific questions.

Also remember to optimize your robots.txt file: don't block LLM bots from accessing your most important content. Check that your site is indexed by Bing (which powers ChatGPT) and by Perplexity.

3. Create content in a question-and-answer format

AI loves clear answers to specific questions. That's why your articles should include FAQ sections, lists, tables, and short definitions. If a client asks “how much does SEO cost,” prepare an answer in the form of a table with price ranges. This increases the chance that AI will choose your content as a source.

Example: instead of writing “our agency offers comprehensive services,” write “typical SEO rates in the region range from a few thousand to several thousand per month, depending on keyword competitiveness. These are market-wide prices - Growth Group mainly charges on a success fee model (a low base fee per service, with the rest tied to results).” Specific numbers and ranges are more valuable to AI.

4. Build links and media mentions

AI often cites sources that are widely linked. Working with media, publishing guest articles, being present on Wikipedia: all of this increases the likelihood of being cited. According to OtterlyAI (2025), being on Wikipedia and working with media increases the chances of being cited in AI answers.

For companies in the region, it's also worth securing local mentions: in regional press, industry portals, and business organizations. This builds authority in the eyes of AI.

5. Monitor traffic from LLMs

In Google Analytics, you can track referral traffic from ChatGPT, Perplexity, and other AI tools. This will show whether your content is being used as a source. If you see increases, it's a sign that optimization is working. If not: analyze which content is being cited by competitors and fill the gaps.

Also remember Link Citation Tracking: monitoring citations and links to your brand in AI answers. You can do this manually by asking questions in ChatGPT and checking whether your company appears.

Measuring the effects of AI optimization

AI optimization is an ongoing process. It is not enough to improve content once: you need to monitor results and adjust your strategy. Here are the key metrics:

• Referral traffic from LLMs: number of visits to your site from ChatGPT, Perplexity, etc.
• Visibility in AI answers: whether your brand appears in answers to industry questions.
• Number of citations: how often AI links to your content.
• Conversions from AI traffic: whether users coming from AI take desired actions (forms, calls).

It is also worth regularly testing how AI answers questions about your industry. Ask questions like "how to choose a marketing agency" and check whether your company appears. If not: analyze who appears and what they do better. An agency owner asks ChatGPT: "which marketing agency is best for B2B companies?" The model lists three competitors with specific examples of their work: his company is not even in the second paragraph. He checks his website: walls of text without headings, no FAQ, no specific numbers.

Remember that optimizing for pure LLMs (without internet access) requires long-term authority building, because the model does not update in real time. That is why regularly publishing valuable content is so important.

Common mistakes in AI optimization

Many marketers make mistakes that make it harder for AI to understand content. Here are the most common ones:

1. Lack of structure: walls of text without headings, lists, and tables are hard for AI to process.
2. Outdated data: AI only cites credible and current sources. If your content contains outdated statistics, you lose authority.
3. Ignoring customer questions: if you do not answer specific questions, AI has no reason to cite you.
4. Lack of monitoring: you do not know whether your content is being used, so you cannot react.
5. Black Hat GEO: attempts to manipulate AI answers (e.g., hidden content) are risky and can harm the brand. Better to focus on ethical practices. A company pastes invisible text on its page with the brand name repeated fifty times, hoping to trick the AI algorithm. Google detects the manipulation, imposes a manual penalty, and the page disappears not only from search results but also from generative answers for months.

In a B2B context, it is also worth remembering that AI often uses content that aligns with evidence-based persuasion principles. According to J. Scott Armstrong (2010), advertising based on empirical evidence is more effective than intuitive practices. Therefore, content should reference research, data, and concrete examples, not just general statements.

AI optimization is not a one-time project but a continuous process. If you need support in this area, B2B marketing strategy can help you build a coherent action plan. In turn, marketing audit will help you assess where you are and what needs improvement.

Remember that B2B competition is growing in the region. Companies that implement AI optimization early will gain an advantage. It is an investment that pays off in the form of visibility in AI answers and real leads.