• Shopware Gold Partner & Shopify Partner

Chapter 4

External authority: visibility beyond your own domain

Table of contents

The first three chapters treated the store itself as the unit of analysis: its technology, its structure, its content. That falls short. An AI system also forms its opinion of a brand from everything credible third parties say about it.

A store can be technically perfect on its own domain and still build much less authority with AI systems if nobody talks about it outside that domain.

An example makes this tangible: a merchant sells high-end espresso machines. On his own website, he calls himself a specialist – as expected, every store does. When Trusted Shops shows thousands of genuine reviews, a coffee magazine tests his products, a YouTube channel compares one of his machines and a manufacturer lists him as a specialist dealer, a different picture emerges: it is no longer only the store claiming expertise – others confirm it.

An Ahrefs study of 75,000 brands (December 2025) shows that brand mentions on the web correlate clearly with AI visibility across ChatGPT, Google AI Mode and AI Overviews – with values between 0.66 and 0.71 depending on the platform. Classic link metrics show a noticeably weaker relationship. Striking: in the same study, YouTube mentions showed the strongest correlation of all factors examined at around 0.74 – even ahead of classic web mentions.

  • 0.66–0.71: correlation of brand mentions with AI visibility, across platforms.

  • 0.74: correlation for YouTube mentions – the strongest factor examined.

  • 84% of all AI citations come from earned media, i.e. unpaid third-party sources.

A complementary evaluation by Muck Rack, as of May 2026, with more than 25 million cited links from ChatGPT, Claude and Gemini across 17 industries, comes to a related result: 84 percent of all citations come from earned media. Paid and advertorial content, by contrast, accounts for only 0.3 percent. Important for context: Muck Rack defines "earned media" broadly; besides classic journalism, it includes academic research, government sources, reference works and content from other companies. Journalism alone accounts for around 27 percent of citations.

Ahrefs itself explicitly points out: this is a correlation, not a proven chain of cause and effect. It does not make backlinks worthless.

What is the cheapest foundation for external authority?

Before more elaborate measures, the cheapest step comes first: keep your brand identity consistent across all platforms. In practice, this means adding the sameAs property to the Organization schema from Chapter 5. Google explicitly recommends sameAs in Organization markup. The property points to external profiles and pages with additional information about the organization and thus supports a consistent attribution of the brand across different platforms – a sensible building block of a consistent digital brand identity, not an independent GEO lever with a strong effect on its own.

What about Wikipedia and Wikidata? Wikipedia and Wikidata can be strong external reference points for a brand's identity. But they cannot simply be used as a marketing channel like a business directory: Wikipedia has strict notability criteria, and entries that read like advertising are removed quickly. For most mid-sized Shopware stores, this is not a realistic first step. Wikidata is more accessible, but it also requires reliable, source-based information rather than promotional claims. For most stores: a possible long-term goal, not a starting point.

How does the arboro trust hub analysis work?

At arboro, we have developed a simple method from this: the trust hub analysis. The starting point: ask ChatGPT, Perplexity and Claude thirty to fifty typical buying advice questions from your industry and document, for each answer, which domains are named as sources.

The trust hub analysis in seven steps:

  1. Collect thirty to fifty typical buying advice questions from your industry.

  2. Ask ChatGPT, Perplexity and Claude these questions and evaluate the answers separately for each system.

  3. Record all cited domains.

  4. Note frequency and topic context for each domain.

  5. Check your own presence on these domains.

  6. Run the same check for your most relevant competitors.

  7. Derive concrete opportunities: on which relevant domains are we missing while competitors already appear there?

One methodological limitation belongs here: the list is not a permanent ranking. Models, sources and answers change. The trust hub analysis is a snapshot that should be repeated at regular intervals.

This shift is now also reflected in the way we work: for SEO and GEO services, we rely less often on classic, isolated link building and more specifically on this trust hub logic instead. The goal is therefore no longer to build as many links as possible. The goal is to appear on the right domains in the right topical context. Ideally, this also creates a high-quality backlink for classic SEO – then both disciplines work together.

What are the four types of external authority?

Your own website remains the foundation of all AI readiness work from Chapters 1 to 3 – but by definition it is not external authority, because the store controls it completely. External authority only arises where someone else talks about the store:

  • Customer authority: reviews and testimonials on platforms such as Trusted Shops. The most direct proof that real customers have bought.

  • Community authority: Reddit, specialized forums, independent discussions. They arise outside your control – and that is exactly why you should not create them artificially, as Chapter 13 shows.

  • Editorial authority: trade media, industry portals, podcasts and independent YouTube channels. Video in particular deserves more attention than it gets in classic SEO work.

  • Partner authority: mentions through business partners. A technical integration alone does not create a mention. Only when the partner actively names you as a reference does the confirmation arise.

Why is this dimension not secondary?

Chapter 3 used the Gemini 3 update to show how quickly AI visibility can shift. That is precisely why external authority is strategically interesting: a model can change its source selection overnight. A presence built up over years in customer reviews, trade media, communities, videos and partner sources, on the other hand, does not disappear with a single model update.

Your own website explains who you are. External authority shows whether others see it the same way.

Our assessment

A store that works exclusively on its own website and ignores external authority is only working on part of the visibility dimension. For most stores, two or three of the four channels mentioned are realistically feasible, not all four at once – more important than completeness is to start with your own trust hub analysis and derive the priorities from it.

What you can check in ten minutes today

  1. Check whether your store's company name, address and description are consistent on the most important external platforms.

  2. Ask ChatGPT, Claude or Perplexity a handful of typical buying advice questions from your industry and use them to start your own trust hub list.

  3. For partner authority specifically, check: are you only a user of the respective systems, or are you also actively named there as a reference?

Frequently asked questions

No. In the Ahrefs study of 75,000 brands, brand mentions do show a stronger correlation with AI visibility than link metrics. But that is a correlation, not a proven chain of cause and effect – and backlinks remain an important signal for classic SEO.

A method for finding out which domains AI systems in your industry rely on. You ask ChatGPT, Perplexity and Claude 30 to 50 typical buying advice questions, record the cited domains and compare your presence there with that of your competitors.

For most mid-sized stores, this is not a realistic first step. Wikipedia has strict notability criteria and removes promotional entries quickly. Wikidata is more accessible, but it also requires source-based information.