Short answer: yes, a growing minority of couples use AI somewhere in planning, and almost every published number about it comes from one company. Longer answer below, including which parts of the evidence I think are weak.
What the industry numbers say
There are two figures in circulation, and it is worth knowing that both come from The Knot Worldwide, which owns The Knot, WeddingWire and WeddingPro.
In a press release dated the third of February 2026, announcing the first wedding industry app built inside ChatGPT, The Knot Worldwide put the figure at thirty six percent of couples actively using AI tools for wedding planning, and described that as nearly double the rate at the beginning of January 2025.
In a WeddingPro report published on the twenty third of June 2026, based on a survey of one thousand four hundred wedding vendors, the figure given was nearly twenty percent of couples planning 2025 weddings reporting that they used AI for planning help.
The more interesting number in that report
Buried in the same WeddingPro survey is the figure that actually describes the situation. Seventy percent of the fourteen hundred vendors said they did not know whether couples were using AI to find them. Only fourteen percent described themselves as very or extremely familiar with using AI in their business, while sixty percent expected it to be essential within five years.
So the industry position in mid 2026 is that most vendors think this will matter, and almost none of them know where they stand today. That gap is the whole reason I started measuring rather than reading surveys.
What a survey cannot tell you, and a measurement can
A survey tells you how many couples say they use AI. It cannot tell you whether AI is naming you. Those are different questions, and only the second one is actionable.
So here is the other half. In September 2026 I measured eighteen wedding businesses in Sweden, Norway, Germany, Belgium and Canada. Twenty to twenty two questions each, written after reading their sites, none containing the business name, three hundred and seventy questions in total. Each question was asked three times to Perplexity, ChatGPT with web search, Gemini and Claude.
35%
of questions where the business held a stable place on at least one of the four services
3 of 18
businesses were named on none of their own questions
1 in 5
of the questions a business appeared on at all, it appeared in only one run out of three
The services disagree with each other more than they agree. Measured as the share of questions where a business held a stable place, meaning it came up in at least two of three runs:
| Service | Questions held stably | Business site cited as a source | Business named in the 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 |
Eighteen wedding businesses, 370 questions, three runs per question per service, September 2026. Gemini returns redirect links rather than a readable source list, which is why its citation count is zero. That is a limit of what can be observed, not evidence that Gemini reads nothing.
Two things follow from that table. Gemini names businesses more than anyone and shows you nothing about why. And Perplexity cites the measured business site in one hundred and ninety eight answers, more than twice as often as any other service, which means your own website can win you a Perplexity answer in a way it cannot win you a Gemini one.
Why one run is not a measurement
Of the questions where a business appeared at all, roughly one in five appeared in only a single run out of three. If you check your own visibility by asking ChatGPT once, you have a one in five chance of recording something that will not be there tomorrow, in either direction.
This is also why the number that matters is the direction across months rather than any single measurement, and why the questions have to stay identical from one month to the next. Change the questions and you have thrown away the only part that means anything.
So what should a wedding vendor actually do
Measure once, properly, so you know where you stand. Then change one or two things and measure again a month later with the same questions. Roughly a month is how long it takes for a new page to be found, read and start appearing in answers, so measuring weekly tells you nothing except that answers move.
Track three things and ignore the rest: the questions where you hold a stable place, the questions where you appeared in exactly one run out of three, and the names that come up in your place on the ones you lose. The middle group is where the least work is left to do, and the third group is your actual competition in this channel, which is often not who you assumed.
And keep the honest frame. In September 2026 this is one channel among several, growing, and cheap to check. The reason it deserves attention out of proportion to its current size is that an AI answer names four or five businesses and then stops. There is no second page to be on.
If you want to see where you stand, the free check asks four questions to ChatGPT and Perplexity and takes about a minute. The full method is written out, and there is a comparison of the tools in this category, including the ones that are not mine.
Questions people ask next
What is AI visibility for wedding vendors?
It is whether your business comes up when a couple asks an AI service an open question, such as who the good wedding photographers in a city are. Two things count as coming up: your name written into the answer, and your website cited as one of the sources the answer was built from.
How often should a wedding vendor check their AI visibility?
Once a month, with the same questions every time. Checking more often tells you nothing extra, because a new page takes roughly a month to be found and read before it can affect an answer. Checking with different questions each time is worse than not checking, because it throws away the only thing worth having, which is the trend.
Which AI service do couples use most for wedding planning?
ChatGPT has by far the largest general audience, and The Knot Worldwide built its wedding app inside ChatGPT rather than anywhere else, which is a reasonable signal about where the industry thinks couples are. But in measurements the four services return very different lists, so being present in one tells you little about the others.
Is it worth worrying about AI if most couples still use Google and Instagram?
It is worth measuring rather than worrying. The honest position in September 2026 is that AI is one channel among several, growing quickly, and cheap to check. What makes it different from Google is that an AI answer names four or five businesses and stops, so there is no second page to be on.