What 41 Million Phone Calls Say About Where Intent Shows Up (and What It Might Mean for Travel)

What 41 Million Phone Calls Say About Where Intent Shows Up (and What It Might Mean for Travel)

Dreaming about a trip and paying for one don’t happen in the same place. Anyone who has spent an evening saving hotel pages and then booked nothing will know that. Call data gives the gap a bit of shape.

Infinity, a call analytics company, has been looking at the AI search shift through more than 41 million inbound calls tracked over a year, for customers in several UK sectors, travel among them. The headline number is that the share of website visits ending in a call went from 1.59% to 2.04% between January 2025 and June 2026. That’s a 31% rise.

Some channels get the phone ringing; most don’t

The full whitepaper splits call behavior by channel, and the gaps are wide. Paid search visitors called on 3.08% of visits and organic search visitors on 1.76%. Display managed 0.46%, and social just 0.18%.

Call length follows a similar pattern. Of paid search calls, 71.1% ran past a minute, as did 73.6% of organic calls and 49.8% of display calls. Infinity is fairly relaxed about display. It reaches a lot of people, produces the odd call, and most of those are short because the person wasn’t ready. That’s what an awareness channel is supposed to do.

What it might mean for travel

None of this is travel data. What follows are hypothetical thoughts, each hung on something in the figures, and none of them tested.

Inspiration might get little credit. If travel behaves like the wider set, the things that make people want a trip (features, social posts, display) might get very little credit in a call report. The call comes later, from somewhere else, probably a search by someone who has already settled on roughly where they’re going. Judge inspiration by the calls it produces directly, and it would look like it does nothing.

Complicated trips might ring more. More than 70% of calls from search and AI referrals in the data lasted longer than a minute, so these aren’t quick queries. A weekend city break is easy to book online. A multi-stop itinerary, a family villa or a honeymoon has more moving parts, and it wouldn’t be surprising if those were the trips where a chatbot’s suggestion ends up as a phone call to settle the details.

The price a chatbot quotes may not survive the booking. In Q2 2026, price and budget mismatch was the leading barrier among sales leads that didn’t convert, with availability second. Infinity explains that buyers research in tools like ChatGPT or Perplexity, pick up an indication of price, and then the number they’re quoted is different. Travel prices move with the date, the room type and what’s included, so the gap between “what I was told” and “what you’re asking” could be wider still. That’s a guess, and a testable one.

Who gets the enquiry might change. The whitepaper says AI tools weigh up options and put one forward, and that the brand it recommends gets the high-intent call. If an AI shortlist names three properties, the fourth may never hear from that traveller. It would be worth knowing whether those shortlists lean toward big brands or whether a small property with strong reviews gets a look-in. The figures can’t say.

The adviser’s job could shift. Infinity’s reading is that callers now arrive better informed. For a travel adviser, that might mean spending less time presenting options and more time checking and correcting ones the guest has already half-chosen. For premium travel in particular, being able to speak to a real person might be part of what the guest is paying for.

Reports could miss phone bookings. The whitepaper argues that a report counting only web form completions measures a shrinking share of conversions. If a travel business reports mainly on online bookings, some of its demand might be finishing on the phone without being counted.

Busy periods could amplify all of it. The published figures don’t separate seasonal patterns. It’s reasonable to wonder whether the weeks when everyone books a summer holiday would intensify each of these effects.

What Infinity thinks is going on

AI search is doing the early work. Buyers ask a chatbot first, arrive on a website better prepared, and ring when they’re ready to commit. The whitepaper sums it up as AI handling discovery and the phone handling commitment.

It points to two jumps in the visit-to-call rate. One, from September to November 2025, lined up with ChatGPT search rolling out more widely. The other, in May and June 2026, set a new baseline above 2.0%. Timing isn’t proof, and the explanation is Infinity’s own.

Checking it against your own records

Anyone with call or enquiry logs could test this without much effort. See what share of web visits end in a call, which channels the calls come from, and whether callers say they’ve already asked an AI tool.

If the pattern in Infinity’s data holds for travel, it should show up there.

One reminder of the limits. The figures are aggregated and anonymized from one provider’s customers, across several UK sectors with no split by sector, so they can’t tell us how travel itself gets booked.