Technical SEO in the AI Era: 11 practical steps to optimize your site now

Technical SEO has always been the foundation of a successful digital strategy. Our job was to give clear signals to search engines: a robots.txt to say where not to go, a sitemap.xml to show the way, and a 301 to indicate a change. It was a game of rules and directives. But the arrival of generative AIs and complex language models (LLMs) has changed the board. Today, following the rules is no longer enough. We need to teach context. The new mission of Technical SEO is twofold: to continue giving clear instructions to crawlers and, at the same time, to provide proof of context to ChatGPT, Gemini, Perplexity, and other chatbots. AI doesn’t just follow a command; it seeks to understand the relationships, entities, and logic behind each signal. In this guide, we’ll detail our conversation with Amanda Noronha about “Technical SEO in the AI Era”, using practical examples to adapt your strategy and optimize your site to be not just crawlable, but perfectly understandable.

The Core Principle: Technical Instructions vs. Semantic Proofs

To understand the changes, think of it this way:

• Classic Optimization (for crawlers): Focused on giving direct instructions. The goal is efficient crawling, correct indexing, and consolidating authority signals.
• Modern Optimization (for AIs): Focused on providing proofs of context. The goal is to eliminate ambiguities, build a robust knowledge graph about your brand, and prove your authority through consistency.

Let’s see how this applies in practice.

1. URLs and Information Architecture: Be a Digital Librarian

A site’s structure is the map that guides both users and robots. The structure of your site is the first lesson you give to AI about your business.

• For Google: The focus was on friendly URLs containing the main keyword. The structure domain.com/category/product was seen as a folder hierarchy that helped Google understand relevance.
• For AI: The keyword in the URL has lost its prominence. AI sees the URL as a unique identifier for an entity (the page). The most important thing is not the URL in isolation, but how it connects to others to form a web of logical information. The site architecture becomes a conceptual map of your business.

Practical Example:

Imagine a tech e-commerce store. The AI doesn’t just read the URL /notebooks/dell/inspiron-15. It understands that Dell is a sub-entity within Notebooks and that Inspiron-15 is a specific product of that brand. The structure of internal links and the organization of categories prove this relationship, teaching the AI your business model.

Practical Action: Think like a Librarian – The analogy is perfect: our job is to organize the site’s information like a librarian organizes a library. It’s not enough to just put books on shelves; you need to categorize them logically so that anyone (or any AI) can find what they need and understand how the topics connect. Good information architecture planning from the start avoids confusion and makes your site an organized source of knowledge.

2. The rel=”canonical” Tag: Consistency is the Proof of Truth

The canonical tag was created to solve duplicate content issues, but AI interprets it with an extra layer of scrutiny.

• For Google: The rel=”canonical” was a clear directive: “Of all these URLs with similar content (e.g., URLs with filters or UTM parameters), consider this one the official.” The crawler generally obeyed.
• For AI: The AI can read the content of all versions, even non-canonical ones, to build its understanding. If the content of the variant page contradicts or differs significantly from the canonical page, it creates inconsistency and dilutes the trust and authority of your main message.

Practical Example:

A product page (/blue-tshirt) has a version with a tracking parameter (/blue-tshirt?ref=promo). The canonical tag points to the first one. However, the version with the parameter displays a different promotional price that is not on the main page. AI detects this inconsistency. It doesn’t know what the real price is, which can weaken trust in the information you provide.

Practical Action: Ensure that essential content (descriptions, prices, specifications) is identical across all variations of a page. The canonical tag is the instruction, but content consistency is the proof.

3. Hreflang Tags: from Translation to Cultural Localization

Hreflang tags are essential for international sites, but their importance goes beyond simple targeting.

• For Google: It was a targeting signal for Google: “Show the /es/ version to users in Spain.” The focus was on correctly directing traffic.
• For AI: AI processes and compares the different language versions. If the translations are low quality, literal, or not culturally adapted, AI notices. This can lead to a weak or incorrect understanding of your brand in that market, affecting the perception of your authority as a global entity.

Practical Action: Don’t just translate, localize. The content must be culturally relevant, using local expressions, currencies, and examples. AI will perceive the quality and authenticity of your communication in each market, which strengthens your reputation.

4. Internal Anchor Text: from Keywords to Concept Networks

Anchor text has always been a pillar of SEO, but its strategic role has evolved.

• For Google: It was a strong relevance signal focused on keywords. Linking to a page with the anchor text “technical SEO services” helped that page rank for that term.
• For AI: Anchor text teaches AI the semantic relationship between two pages. Always using the same anchor text limits this learning. By varying the text, you create a much richer conceptual map.

Practical Action: Vary the anchor text with synonyms and contextual phrases. Instead of just using “technical SEO services” also use “our technical SEO approach”, “Core Web Vitals optimization”, or “see how a technical audit works.” This teaches AI the entire network of concepts surrounding your main service, proving the depth of your specialization.

5. Sitemap.xml: from a List of URLs to an Organized Map

The sitemap has always been crucial, but how we structure it now makes all the difference.

• For Google: The common practice was to generate a single sitemap with thousands of URLs and expect Google to figure it out to find and index everything.
• For AI: A giant, disorganized sitemap makes the job difficult for all crawlers. AI, just like Google, benefits from an organized structure. Dividing the sitemap into smaller, logical files (a sitemap index) speeds up crawling and helps AI understand the different sections of your site.

Practical Example:

Instead of a sitemap.xml with 50,000 URLs, create an index that points to specific sitemaps:

• sitemap-pages.xml
• sitemap-blog-posts.xml
• sitemap-products-category-a.xml
• sitemap-products-category-b.xml

This organization not only facilitates indexing but also helps you diagnose coverage problems in Google Search Console much more accurately.

6. Structured Data (Schema.org): The Official Language of AI

If there’s one element that has become exponentially more important, it’s structured data.

• For Google: Schema was used primarily for a tactical goal: getting rich results (review stars, FAQs, recipe prep time) in the SERP.
• For AIs: Schema.org is the fundamental language for building your brand’s knowledge graph. AI consumes this data to understand who you are, what you do, and how your different parts connect, even if there is no associated rich result.

Practical Example:

Your blog publishes a chocolate cake recipe.

• For Google: You use the Recipe schema so the recipe appears with reviews and prep time in the results. The focus is on display.
• For AI: You go further. In addition to Recipe, you add:
  • Article for the post itself.
  • Person for the author, using the sameAs property to link their profiles on LinkedIn or Twitter.
  • Organization on your About Us page, using sameAs to connect the company to its Wikipedia page, social profiles, and Google Business Profile.

AI uses this data mesh to understand that the recipe was created by a real person, who is part of a legitimate and reliable organization. This drastically strengthens the pillars of E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness).

7. 301 Redirects: The instruction is not enough, tell the story of the change

The 301 redirect is the perfect example of the paradigm shift.

• Before: A domain or URL change was resolved with a 301 redirect. It transferred authority (link equity) and instructed Google to index the new page. Problem solved.
• Now (in the AI Era): The 301 is just the technical instruction. AI, which has learned from all the web’s content, may still see the old brand and the new brand as two separate entities. You need to communicate the change on all fronts to merge the knowledge.

Practical Action: The Case of the Rebranding Task Force

An isolated redirect is not enough. You need to create an ecosystem of signals that tell the story of the change to users and to AI.

1. Implement the 301: This is still the essential technical instruction for crawlers.
2. Create an explanatory page: Develop a page or a blog post detailing the transition. Example: [Old Brand] is now part of the [New Brand] group.
3. Communicate proactively: Use pop-ups on the homepage, banners, and posts on all social networks to announce the change. The goal is for the user to have no doubts about the transition.
4. Use structured data: In the new site’s Organization schema, use the alternateName property to declare the old name, formally connecting the two entities for AI.
5. Validate the understanding: After the change, search Google for the old name. The ideal result is for Google to already show the new site in the results, indicating that it has understood the correlation. This proves that your task force worked.

8. JavaScript and Rendering: Content needs to be accessible to everyone

The discussion about JavaScript and SEO has gained a new layer of urgency. The question is no longer just whether Google can render it, but which AI systems can.

• For Google: The concern was that Googlebot would enter a rendering queue to process JS, which could delay indexing. The solution was to focus on Server-Side Rendering (SSR).
• For AI: The scenario is more complex. While Google (Gemini) and Apple (Applebot) render JavaScript, many other AI crawlers, such as those from ChatGPT, Perplexity, and Meta, do not. If your main content depends on Client-Side Rendering, it will be simply invisible to these systems.

Practical Action: Test, Validate, and Guarantee the HTML

The essential content of your site needs to be present in the raw HTML, before any JavaScript execution.

1. Difference between JavaScript and JSON-LD: It’s a common mistake to confuse the two. JSON-LD is a structured data format (Schema) that can be inserted in a script, but it is not functional JavaScript. The main content of your page should not depend on the execution of complex JS frameworks to be displayed.
2. Test with Different User Agents: Rendering can vary drastically between crawlers. Tools like Prerender.com or TechnicalSEO.com (Mobile-Friendly Test) allow you to simulate how different agents (Googlebot, Bingbot, and even AI crawlers) see your page. This helps identify if your title, description, or main content is being rendered correctly for everyone.
3. Use Google Search Console: The Inspect URL tool is your ally. Check the HTML tab to see the code Google accesses and the Screenshot to confirm if the page is rendered visually. If you see a blank page or missing content, it’s a red flag.
4. Prioritize SSR or SSG: To ensure universal access, use Server-Side Rendering (SSR) or Static Site Generation (SSG). These approaches deliver the main content in pure HTML, ensuring that all bots, from the simplest to the most advanced, can read your information.

If an AI crawler cannot read your content, it may try to infer information from your structured data. However, this is a stopgap. The source of truth should always be the rendered HTML. Recommended Reading: Check out the OpenAI guide: ChatGPT and AI crawlers can read JavaScript: What you need to know.

9. The Disavow Tool: from cleaning links to building authority

The link disavow tool has always been reactive. Now, the strategy needs to be proactive.

• For Google: It was a tool to tell Google: “Please ignore these low-quality backlinks. Don’t count them in the authority calculation.” The goal was to clean up a toxic link profile.
• For AI: AI, on its own, does not obey the Disavow file. It reads the web as it is. If your brand (Entity Y) is mentioned massively on spam sites (Entity X), AI can create a negative association between them, regardless of the disavow.

Practical Action: The Disavow tool is still important for Google’s ranking algorithm. However, the main strategy must be the proactive building of authority. Generate links and mentions on high-quality sources that reinforce the correct associations. The goal is to drown out the negative noise with positive and clear signals, teaching AI that your brand is associated with reliable entities.

10. Content Behind Clicks (Tabs and Accordions): the clarity of visible information

The way we display content on the page directly influences its interpretation by machines.

• For Google: The discussion was whether hidden content in tabs or accordions received the same ranking weight. Over time, Google confirmed that it could render and index this content normally, making the practice acceptable from a technical point of view.
• For AI: AI is not just indexing; it is synthesizing answers. Content that is immediately visible in the HTML is processed as a priority. If crucial information is hidden behind a click, AI may find it, but it may not give it due weight when building a response, considering it secondary. Immediate visibility is proof of importance.

Practical Action: Prioritize Essential Information – Do not hide the answers to your users’ most important questions. If a paragraph is fundamental to understanding your product or service, leave it visible by default. Use tabs and accordions for complementary or low-priority information, but never for the main content. Clarity for AI begins with visual clarity.

11. Speed and Core Web Vitals: from user experience to crawler efficiency

Site speed has always been important, but the reason for it has expanded.

• For Google: Core Web Vitals (CWV) were introduced as a ranking factor focused on user experience. A fast, stable site pleases visitors, and Google rewards this with better positioning. The goal was to optimize for people.
• For AI: AI depends on massive and constant crawling of the web to train its models. A slow, unstable site with loading errors consumes more computational resources (crawl budget) from companies like OpenAI, Google, and Perplexity. In practice, an inefficient site becomes expensive for AI to visit, which can result in less frequent or incomplete crawling of your content.

Practical Action: Optimize for the Machine and the Human – Optimizing Core Web Vitals is no longer just about UX. Think of it as a way to facilitate the work of AI. By ensuring your site loads quickly (good LCP), is interactive (good INP), and stable (good CLS), you not only improve the user experience but also become an efficient and low-cost source of information for all bots, ensuring your content is processed completely and regularly.

Conclusion: from architect of signals to curator of context

Your job as a Technical SEO professional is evolving. The mindset needs to change.

1. Audit with two lenses: When conducting an audit, don’t just ask yourself, Is this following Google’s rules? Also ask: Would an AI reading this code and content understand my intent unequivocally?
2. Prioritize clarity over cleverness: Many old SEO tactics were about being clever to manipulate the algorithm. Now, the winning strategy is to be extremely clear. Ambiguity is your greatest enemy.
3. Connect the technical to the content: Technical work doesn’t end in the code. It must ensure that the story the content tells is the same one the technical structure supports. Your function is to ensure there are no contradictions between what you say and what you signal.

FAQ

1. What is Technical SEO? Technical SEO refers to all the optimizations on a site’s infrastructure that facilitate the crawling, indexing, and comprehension of content by search engines and other technologies, such as AI.
2. Does AI make old Technical SEO practices obsolete? No. Practices like 301 redirects, sitemaps, and canonical tags remain fundamental as instructions for crawlers. The change is that they now need to be reinforced with clear and consistent content to prove the context to the AI.
3. Why is Structured Data (Schema.org) so important for AI? Because it translates your content into a language that machines understand natively. This allows AI to organize the information on your site into a knowledge graph, understanding entities, attributes, and relationships, which strengthens your authority.
4. How does URL structure impact AI? AI uses URL structure and internal links to understand the information architecture and business model of your site. A logical structure helps AI map how different concepts and pages relate to each other.
5. What is the main mindset shift for the Technical SEO professional? The main shift is moving from a compliance with rules mindset to one of clarity and consistency. The ultimate goal is to eliminate any ambiguity, ensuring that the technical signals and the content tell the exact same story.

Check out “Technical SEO in the AI Era,” the second episode of the No-Nonsense Technical SEO program: