AI search does not make technical SEO obsolete. If a search system cannot discover, access or understand an important page, the quality of its writing may never matter. For Google AI Overviews and AI Mode, the starting point is straightforward: eligible supporting pages must be indexed and eligible to appear in Google Search with a snippet. ChatGPT search has its own discovery controls. That makes technical accessibility the first thing to check, not an optional extra.
This checklist is for a business website, not a large technical publishing platform. It separates what can be verified from what cannot be guaranteed.
1. Check whether your important pages can be discovered
Create a shortlist of your homepage, primary service pages and best articles. For each URL, verify that it appears in the website's navigation or is linked from another relevant page. An XML sitemap helps discovery, but it is not a substitute for working internal links.
Open your sitemap and compare it with the shortlist. Remove redirects, unpublished URLs and genuine error pages from the list of canonical pages. When you publish a new article, its URL should appear in the sitemap only once it is public and indexable.
An overlooked issue is a beautiful page that exists technically but can only be reached by typing its address. That page needs a useful contextual link.
2. Inspect crawl access and indexing
Use Google Search Console's URL Inspection tool to inspect an important URL. Check the indexing status, selected canonical URL and whether the live page is accessible. If a page is absent, investigate the reason rather than repeatedly requesting indexing.
Check the HTTP response and page instructions:
- A normal published page should return a successful response and should not accidentally contain
noindex. - A removed page should use an appropriate redirect or genuine not-found response, not a misleading success page.
robots.txtshould not block Googlebot from content you want available in Google Search.- Snippet restrictions such as
nosnippetcan affect eligibility for Google's AI search links.
An indexable page is not necessarily an indexed page. Google still decides which pages it includes.
3. Understand which AI crawlers matter
For ChatGPT search, OpenAI documents OAI-SearchBot as the crawler associated with search discovery. GPTBot has a different purpose related to potential model training. Allowing one does not automatically require allowing the other. ChatGPT-User is associated with some user-triggered visits and is not a substitute for search crawling.
Check robots.txt, any firewall rules and, where available, verified server logs. A user-agent name alone is not proof that a request came from a genuine crawler. Avoid blocking legitimate traffic through overly aggressive security settings.
Do not promise that granting crawler access will produce ChatGPT citations. It simply removes one potential barrier.
4. Keep the actual answer accessible as text
Search engines and AI retrieval systems benefit from meaningful information in readable page content. Put your service name, audience, location and main explanation in real HTML text instead of hiding everything in an animation, an image or an interactive widget.
For JavaScript-heavy websites, inspect what the initial response and the rendered page actually contain. Test the live URL, not only the development preview. Important content should not require someone to click through several controls before it becomes understandable.
A practical test: disable the visual design mentally and read just the text and headings. Can someone still identify what the company does and why the page exists?
5. Make canonicals and language versions consistent
Each live English or Dutch article should have its own self-referencing canonical URL. Equivalent language versions should reference each other with reciprocal hreflang annotations. A translated Dutch page should not point its canonical at its English twin: they serve different audiences.
Keep article dates truthful. If an older article is substantially revised, update its modification date, not its original publication date. Use structured data such as BlogPosting only when the values match the visible page.
6. Check usability before chasing technical scores
A page should load reliably, work on a phone and allow visitors to navigate to related services. Watch for content hidden under mobile menus, links that cannot be tapped and consent dialogs that cover the page. Core Web Vitals can reveal real problems, but improving a score does not guarantee a higher ranking.
A 30-minute technical SEO audit
| Time | Action | Evidence to capture |
|---|---|---|
| 5 minutes | Open robots.txt and sitemap | Access rules and public URLs |
| 10 minutes | Inspect three important URLs in Search Console | Indexing and canonical status |
| 5 minutes | Follow internal links to those pages | Whether they are discoverable |
| 5 minutes | Check mobile rendering and text visibility | Readability and working links |
| 5 minutes | Review EN/NL tags and Article data | Accurate page metadata |
Record each issue with a URL, evidence, owner and priority. That turns the checklist into a fixable plan instead of a vague SEO score.
Does technical SEO guarantee visibility in AI answers?
No. Google explicitly says that there are no additional technical requirements for AI Overviews or AI Mode beyond its applicable Search requirements, and that meeting them does not guarantee inclusion. Technical SEO establishes eligibility. Original, relevant content and a trustworthy website are still essential.
If you are planning an AI visibility program, begin with an accessible website, then develop useful content around customer questions. Read our AI visibility measurement guide and explore AEO DIRECT's SEO services.



