Is Your TYPO3 Website Ready for ChatGPT?

Classic search engine optimization is changing. While it used to be mainly about ranking as high as possible on Google, ChatGPT, Google AI Overviews, Perplexity, and other AI systems now form an additional gateway to information on the web.

This raises a new question for companies:

This is exactly where technical optimization for AI systems begins. Many measures overlap with classic technical SEO. Others gain significantly in importance because of AI search.

For TYPO3 websites, three levels can be distinguished.

1. Can the AI Read the TYPO3 Website?

The first requirement sounds obvious: a website's content must be accessible to machines in the first place.

Modern AI systems access web content in different ways. Some information comes from search engine indexes, other systems use their own crawlers or access current web pages via search services.

Websites where essential content is only generated via JavaScript, or whose HTML structure is unnecessarily complicated, can therefore be technically problematic.

TYPO3 generally brings good conditions for this. Content can be delivered server-side as complete HTML.

Among the decisive factors are:

The last point in particular is becoming increasingly interesting.

robots.txt: Which AI Is Allowed on Your Website?

Alongside classic search engine crawlers like Googlebot, a wide range of specialized AI crawlers now exists.

Companies should therefore consciously decide how to handle crawlers from OpenAI, Anthropic, and other AI providers, for example.

It's important to distinguish what the access is for. Fetching a website for a live AI answer, for instance, is not necessarily the same as using content to train an AI model.

A robots.txt should therefore no longer just be set up once and then forgotten.

Controlling AI crawlers is becoming part of a modern technical website strategy.

2. Can the AI Understand What the Page Is About?

A website can be technically perfectly accessible and still be difficult for an AI to interpret.

Let's take a typical TYPO3 corporate website.

It might contain:

A human quickly understands the connections. For a machine, however, these relationships need to be recognizable as unambiguously as possible.

Is WACON a company? Is TYPO3 a product or a service? Is Frankfurt the company's headquarters? Is the named person the author of an article, a point of contact, or the managing director?

This is exactly where structured data and Schema.org come in.

Schema.org Is Becoming More Important for AI Visibility

Structured data can be used to unambiguously mark up a website's information.

In TYPO3, for example, the following can be stored or output automatically:

Organization

• Company name
• Logo
• Address
• Website
• Contact details
• Social media profiles

Person

• Name
• Position
• Company
• Profile

Article

• Headline
• Author
• Publication date
• Modification date
• Topic area

Service

• Service offered
• Provider
• Service area

BreadcrumbList

• A page's position within the website structure

Depending on the website, further types such as Product, Event, JobPosting, or LocalBusiness may be added.

Technically, this information is often integrated into the source code via JSON-LD.

The advantage:

The machine no longer has to infer what a piece of information means purely from body text. The website delivers the meaning directly.

This is helpful for classic search engines – and becomes even more important in the context of generative AI.

From SEO to "Meaning"

In classic SEO, the conversation for a long time was about whether Google can crawl and index a page.

With AI systems, another layer is added:

For companies, this can be decisive, for example, when users ask questions like:

To do this, an AI needs to be able to connect companies, services, locations, references, and areas of expertise with each other.

The more clearly a website communicates these connections, the less an AI system needs to infer them itself from scattered pieces of text.

3. Can an AI Interact With the Website?

The third level goes significantly further still.

Today, we're mostly talking about whether ChatGPT or another AI finds information on a website and uses it in an answer.

In the future, however, AI agents will increasingly carry out actions.

A user might say, for example:

This means it's no longer enough to merely display information on an HTML page.

Websites and applications increasingly need structured interfaces through which machines can retrieve information and trigger actions.

APIs Are Becoming More Important for Websites

Many TYPO3 projects already have interfaces today.

TYPO3, for example, can provide content to other applications via APIs. This is already common in headless architectures.

In the context of AI agents, this architecture gains additional importance.

Instead of an AI system interpreting a web page the way a human would, it could retrieve structured information directly:

Standardized protocols for communication between AI systems and applications could play a special role here.

One example is the Model Context Protocol (MCP).

Standards like this could make it easier for AI agents to address external systems in a structured way in the future.

This means the website is developing, over the long term, from a pure information surface for humans into an information and service platform for both humans and machines.

What About llms.txt?

The llms.txt file is frequently discussed in the context of AI optimization.

The idea is initially reminiscent of robots.txt: website operators provide a file specifically intended for language models, meant to make a website's important content and resources easier to find.

The concept is interesting.

However, its importance shouldn't be overestimated at this point.

An llms.txt replaces neither:

Anyone looking to optimize their website for AI systems should therefore first improve the technical foundations.

An llms.txt can then be a sensible addition.

TYPO3 Has Good Conditions for AI-Optimized Websites

A structured content management system like TYPO3 in particular has a decisive advantage.

Much of the information already exists in structured form.

TYPO3 knows, for example:

The decisive task is to make this existing structure machine-readable to the outside world as well.

A piece of information should therefore not only appear visually on a web page.

It should, as far as possible, also be described unambiguously on a technical level.

The TYPO3 AI Readiness Check

Companies with an existing TYPO3 website can already check today how well their website is prepared for AI search.

In particular, the following areas should be examined:

Conclusion: AI Optimization Starts With the Technical Foundation

Optimizing for AI search systems is not a completely new discipline.

We already know many of the fundamentals from technical SEO, accessibility, and clean web development.

The difference increasingly lies in who a website is technically built for.