Is AI Moving Demand to the Phone? What 41 Million Cross-Sector Calls Suggest for Retail

Infinity, a UK call analytics provider, argues that it is. In a recent post built on over 41 million inbound calls recorded across 12 months, the company says demand is shifting towards the phone rather than vanishing, as buyers research and compare with assistants like ChatGPT, Perplexity and Gemini. Its data is consistent with that view, but it cannot show that AI is the cause.

What 41 million calls show

Across the websites Infinity tracks, 2.04% of visitors were choosing to call by June 2026, against 1.59% in January 2025, and calls logged from January to June 2026 were 24% higher than in the same months of 2025. Daniel Wilkinson, Infinity’s Chief Customer Officer, puts it this way: “While traffic might be thinning, demand is not disappearing.” The whitepaper itself does not report a trend in overall website traffic.

Timing is the main hint at a cause. Infinity reports that the share of visitors choosing to call jumped 27% over three months starting in September 2025, which the whitepaper says coincided with the broader rollout of ChatGPT search capabilities. Infinity’s own press summary says that rise, and a second one in May–June 2026, came as AI search adoption accelerated, though the whitepaper names a specific event only for the first. Overlap in timing is not the same as cause.

On these numbers, calls directly attributed to AI referrals are also too few to explain the rise on their own. Over the last year, Infinity attributed 16,381 calls to visitors who clicked through from an AI platform, against 41 million calls overall. If AI is lifting call volumes, it is doing so largely through routes this attribution cannot see, and the data cannot show that. 

The AI-referred calls themselves scored 1.45% on Infinity’s Smart Match event rate, its measure of offline conversion events, against 0.99% for paid search, though from a small base. Infinity’s separate data on AI referral traffic stops in April 2025, so it cannot show how many visits the channel sends today.

What “cross-sector” does and does not mean

Infinity says its call tracking covers sectors including retail, and describes the whitepaper’s data as coming from every major sector. The whitepaper itself publishes no retail breakdown, so these figures are a cross-sector signal for retailers to test against their own records, not a retail benchmark. The summary Infinity published says the calls were tracked in UK sectors, retail among them, and the whitepaper provides no market-level split either.

Price and availability: the two biggest barriers to lost sales leads

Infinity’s Smart Outcomes tool classifies call outcomes, including why sales leads fail to convert. Price and budget mismatch led that list in Q2 2026, at 31.6%. Availability came next at 18.3%, having hardly appeared in 2025, and a new category called product not offered came in at 12.8%. A year earlier, the picture was different: decision-making led at 20.8%, with pricing uncertainty (17.8%) and product understanding (17.7%) close behind. Product understanding as a barrier has since fallen to roughly 8%. In the installations where the tool is used, roughly three in four calls verified as carrying buying intent still ended without a positive result.

Infinity reads this as buyers consulting ChatGPT or Perplexity first, and in many cases coming away with an indication of price. Its view is that when the agent’s quote differs from what the assistant described, the call can fail even though the lead is genuine. One caution: Infinity says the labels on its barrier categories were altered during a taxonomy update, so the broad direction is more dependable than exact percentage-point shifts.

For retailers, price and availability are things that can be checked directly, which makes this a comparatively easy finding to test. Infinity’s post also says that, for many considered purchases, the decision to commit is made on a call. Retail categories built on larger or more complicated decisions may be where that matters most, though Infinity does not say which categories those are.

Checks worth making before peak trading

Infinity’s recommendations include preparing phone teams for price objections, treating AI referrals as a source in their own right, and assigning someone to each stage from ad spend to closed sale. The checks below are suggestions based on the findings, not retail guidance from Infinity.

• Ask the assistants. Put the questions a shopper would to ChatGPT, Perplexity and Gemini about your key products, prices, ranges and stock, and note what comes back. Answers can vary by assistant and over time, so treat it as a spot check.
• Compare with reality. Check those answers against live prices, current stock levels and what your stores and contact teams would actually say.
• Brief the people who answer. A caller who arrives holding a figure needs a team that knows the current price position and can explain it.

Limits worth knowing before you act

All of this comes from one provider’s customers, who are businesses that have chosen to track their calls, so the sample may lean towards businesses where the phone matters. As a call analytics business, Infinity sees what happens around tracked phone calls, which makes it a single viewpoint rather than the full picture. 

The whitepaper does not say whether the rise in call volume is adjusted for changes in Infinity’s own customer base. The idea that buyers research with AI before they ring is Infinity’s interpretation, and the data says nothing about in-store sales, online baskets or how shoppers compare prices elsewhere.

A simple first step is available to any retailer that logs its calls or enquiries. Take last month’s lost sales, tally how many were due to price or stock, and see whether the pattern looks like the one Infinity describes.