This article explains why good rankings and AI visibility come apart, where you would have to be visible today, and how to close the blind spot. It is about the channel logic, independent of platform; the technical work inside your own shop system is a topic of its own.
Visibility is no longer one channel, it is several
A few years ago “visibility on the web” was practically identical with “Google ranking”. Whoever stood at the top was found. Today the same question splits into several separate channels, and a prospective buyer uses them differently:
- Classic Google search: the ten organic results that SEO has always revolved around.
- Google AI Overviews: the AI summaries above the actual results, which give a direct answer and name only selected sources.
- Generative AI systems: ChatGPT, Perplexity and others, which customers increasingly consult before they open a search engine at all.
These channels draw on different sources, weigh them differently and name different brands. A top position in one is no guarantee of presence in another. Whoever watches only the classic Google position now sees just a section of the picture.
Search engines weigh signals, language models process meaning
The core of the problem lies in how a search engine and a language model read. A search engine weighs relevance signals: matching search terms, backlinks, technical hygiene, user behaviour. A language model, by contrast, processes meaning. It needs content that makes clear what a brand stands for, which category it belongs to and where its difference lies.
Two constellations follow from this, and both occur constantly in practice:
Top ranking, but no AI mention. A shop sits at the very top on Google because its pages are technically clean and optimised for the right terms. The texts, however, are promotional, interchangeable and say little that is concrete. For the search engine that is enough. For a language model that is supposed to write a recommendation there is nothing usable in it, so the brand drops out of the AI answer.
AI mention despite a moderate ranking. The other way round, a shop is recommended by ChatGPT although its own website ranks only averagely. The reason: the brand appears in many independent comparisons, guides and reviews which state clearly what it is good for. The language model has a sharp picture, and names it.
The consequence is uncomfortable: your investment in SEO does not automatically pay into your AI visibility. They are related but separate tasks. Whoever ignores that keeps optimising one channel and quietly loses another.
Google AI Overviews, the channel right on your doorstep
One special case deserves particular attention, because it sits exactly where you used to feel safe: in the Google results themselves. For many queries Google places an AI-generated summary above the classic results. It answers the question directly and points to a few sources.
For shops that shifts the rules on familiar ground. You can still be recognised at position one and still disappear underneath the AI answer the user reads first. If your brand is named in that summary, you benefit twice. If it is passed over, your actual visibility falls although your ranking position still looks “good”. The old metric suggests a safety that no longer exists.
Where you would have to be visible today
To summarise, these are the channels that count for an e-commerce shop in the German-speaking market today, and what matters in each:
| Channel | What counts | What good rankings alone do not cover |
|---|---|---|
| Classic Google search | Position, clicks | Covers this one channel only |
| Google AI Overviews | Mention in the summary | Position 1 does not prevent being passed over |
| ChatGPT, Perplexity and others | Mention in the answer | Different sources, different result |
Three channels, three logics. A single metric, the Google position, can no longer represent that. Whoever watches only that one flies with half an instrument panel. Which AI metrics Rivly records for this is shown on the AI visibility page.
An example that makes the gap visible
A shop for outdoor equipment from the German-speaking market ranks second on Google for its central category. The owner is satisfied. Then he runs the counter-check on the other channels.
In the Google AI Overview for the same category his brand is not mentioned: two competitors and one large marketplace are named. Asked “which shops for outdoor equipment do you recommend in Germany?”, ChatGPT names three brands, and his is not among them. Perplexity at least names it in fourth place. The conclusion: in the channel he measures, he is strong. In the two he has never checked, he is weak to invisible, and that is exactly where a growing share of buying journeys begins today.
The value of this exercise lies not in the alarm but in the clarity. Only the view across all channels shows where the real work lies. Without it the owner would have kept investing in a channel that already runs, and missed the one that decides future growth.
How to close the blind spot
The path is simple in principle, but laborious when you walk it by hand:
- Define realistic questions: the ones your customers would ask without knowing your brand.
- Check every relevant channel: classic Google position, Google AI Overview, ChatGPT and Perplexity.
- Compare with the competitors: not “are we mentioned?”, but “who is mentioned instead of us?”.
- Repeat regularly: because all three channels change, and only repetition shows whether your work has an effect.