What was measured
Eighteen wedding businesses, thirteen in Sweden, two in Norway, one each in Germany, Belgium and Canada. Twelve measured on the Swedish market, two on the Norwegian, four on the English speaking market.
Each business had twenty to twenty two questions written for it after reading its website, none of them containing the business name. Three hundred and seventy questions in total. Every question was asked three separate times to Perplexity, ChatGPT with web search, Gemini and Claude, between the first and fifth of September 2026. That is four thousand four hundred and forty answers, all stored.
A business holds a question when it is named in the answer, or its site is cited as a source, in at least two of the three runs. Appearing in exactly one run of three is counted separately and called a near win. The full method is on the method page.
Headline findings
35%
of all questions were held stably by the business they were written for
3 of 18
businesses were not named on a single one of their own questions
1 in 5
appearances happened in only one run of three, so a single check would have misread them
1. Most wedding businesses lose most of their own questions
Across the eighteen, the median business held about a third of its own questions on at least one service. Five businesses held more than half. Eight held fewer than one in five. Three held none at all.
These are questions written specifically for that business, about its own city, its own style and the venues it actually works at. The questions were not hard.
2. The four services disagree with each other
| Service | Questions held stably | Business site cited | Business named in text |
|---|---|---|---|
| Gemini | 21.1 percent | 0 answers | 239 answers |
| Perplexity | 18.4 percent | 198 answers | 189 answers |
| ChatGPT with web search | 14.9 percent | 69 answers | 170 answers |
| Claude | 10.5 percent | 93 answers | 112 answers |
Share is of 370 questions per service. Gemini returns redirect links rather than a readable source list, so its citation count cannot be observed and is recorded as zero rather than estimated.
Gemini names businesses most often and shows least about why. Perplexity cites the measured business site in one hundred and ninety eight answers, more than twice as often as any other service, which means a business site can win a Perplexity answer in a way it cannot win a Gemini one. Claude is the hardest of the four to enter.
3. The sources are directories and competitors, not magazines
Counted across six full photographer measurements in Sweden, Germany and Belgium, the sites cited most often were photographer directories.
| Source cited | Times cited across six measurements |
|---|---|
| mywed.com | 277 |
| wezoree.com | 204 |
| wpja.com | 150 |
| thewed.com | 130 |
| reddit.com | 93 |
| fearlessphotographers.com | 68 |
Six photographer measurements, Sweden, Germany and Belgium, September 2026. After these, the next most cited sources were other photographers own websites.
Wedding magazines and blogs barely register. In the European measurements neither The Knot nor WeddingWire appears among the top sources, although both dominate when the same kind of question is asked about the American market. Which directory matters is local, and getting that wrong is an expensive way to stay invisible.
4. Broad geography is much harder than one city
The three businesses positioned across several countries rather than one city held nought, five and fifteen percent of their questions, the three weakest results in the set. Three businesses is far too small a number to draw a rule from, and this should be read as something to watch in the next edition rather than a finding.
City size explained less than expected. The three strongest results were outside the capital, but the spread inside single cities was wide enough that on eighteen businesses, city size does not explain much.
What this does not show
It does not show what any individual couple saw, since answers vary by person, phrasing and moment. It is not a random sample of the wedding industry: these are businesses that agreed to be measured, most of them Nordic, most of them photographers. It says nothing about how many couples use AI, only about what the services answer when asked. And it is one month, so there is no trend in it yet. The October edition will have the first one.
Citing this
Free to quote, in full or in part, with a link to this page. If you want a figure that is not published here, or the breakdown for a particular market, ask and I will run it if the data supports it and say so if it does not.
Suggested citation: Lundin, K. (2026). Wedding AI Visibility Index, September 2026. Wedding Mentions. https://weddingmentions.com/data/wedding-ai-visibility-september-2026