Imagine a customer who asks a chatbot what something should cost and whether it is in stock, then rings the business to buy. If enough customers now behave that way, the call that follows is a different kind of conversation from one placed by someone still working out what to buy.
Call tracking platform Infinity says that shift has already happened. Its published view, drawn from a whitepaper and more than 41 million calls logged over 12 months, holds that buyers research in chatbots and then phone to commit. The data points that way, but it records outcomes rather than motives, so the question in the headline is better treated as open.
More people are ringing
By June 2026, 2.04% of the visitors to sites Infinity monitors were phoning, against 1.59% at the start of 2025. Across the calls it tracks, the opening half of 2026 brought 24% more than the equivalent stretch of 2025. Infinity's Daniel Wilkinson, who holds the Chief Customer Officer role, puts it this way: "while traffic might be thinning, demand is not disappearing." Volume alone says nothing about why people ring, and Infinity's published figures contain no time series for overall website visits.
Timing offers a hint. Infinity reports that visitors phoned 27% more often across the three months beginning September 2025, a period the whitepaper describes as coinciding with ChatGPT's search capabilities being rolled out more widely, and the rate stepped up again in May–June 2026.
A separate summary written by Infinity says both rises came while AI search adoption was accelerating. Matching dates are not proof of cause. Calls traced straight to an AI platform totalled 16,381 over the year, against 41 million overall, so if chatbots explain much of the rise, most of that influence must sit outside what this tracking can see.
Are callers ringing to learn, or to settle terms?
The closest the data comes to the question is Smart Outcomes, Infinity's call-classification tool, which sorts calls by intent and, for failed sales leads, by what got in the way. Across the year, it confirmed 1.3 million of the 5.75 million calls it processed, about 23%, as carrying purchase intent, and roughly three-quarters of those never produced a positive outcome.
For failed leads, it also records the barrier involved, and the percentages below are shares of those failed leads. In Q2 2025, the top barrier was decision-making (20.8%). Pricing uncertainty came second (17.8%), and product understanding third (17.7%). By Q2 2026, price-and-budget mismatch had taken the top spot at 31.6%. Availability, almost absent in 2025, had reached 18.3%, a new "product not offered" category stood at 12.8%, and product understanding had fallen to about 8%.
Put plainly, a smaller share of failed sales leads now come down to product understanding, and larger shares come down to price and availability. The whitepaper's account is that buyers now research in assistants like ChatGPT or Perplexity and often come away with a price indication, so a quote that differs from it can sink the call although the lead was genuine. Infinity's post reports Wilkinson's view that more callers now ring expecting to be closed while call scripts and sales staff remain set up to educate.
Both are readings, not measurements. The data shows why leads failed, not why customers picked up the phone, and Infinity cautions that the names of its barrier categories changed when it switched classification scheme, so the general pattern is a safer guide than any exact percentage-point shift.
How much weight can this evidence bear?
Less than a confident answer would need. The barrier shares describe failed sales leads, so they say nothing about the calls Smart Outcomes did not verify as carrying purchase intent, which were most of the calls it processed over the 12 months. All the figures come from one company's clients, who signed up to track calls; that may skew the sample towards firms that live by the phone.
Infinity says the study spans every major sector, and names automotive, retail, legal, healthcare, travel, home services and financial services among those it tracks, but the whitepaper has no sector or country breakdown, and Infinity's own summary refers to UK sectors. The paper is also silent on whether the 24% rise reflects any growth in Infinity's own client list. And nothing in the data shows what any chatbot told a customer about price or stock.
If callers have changed, what should businesses do?
Infinity's own advice includes briefing call handlers on price objections and moving scripts from education towards closing, giving AI referrals a line of their own in the reporting, and using what callers ask about to shape FAQs, messaging and creative. Those are recommendations, and neither the whitepaper nor Infinity's post presents test results for them. Three modest checks, offered as suggestions rather than Infinity's advice:
- Where you record calls lawfully, sample recent ones and an earlier batch, and sort each by why the customer rang: to buy, to check price or stock, to get help, or something else. Comparing the two gives a rough view of whether your own mix has shifted, instead of borrowing anyone else's.
- Ask a few chatbots the questions a customer might ask about your products, prices and availability, and set the answers beside what your staff would genuinely quote and deliver. Answers can differ by tool and over time.
- Check whether your call scripts and training assume a caller who needs educating or one who is ready to settle terms.
The headline question has no firm answer yet, but any business that keeps records of its calls can at least check whether its own mix of callers has shifted.
