According to a reference report published by Invoca on July 13Calls referred by ChatGPT are more likely to be considered leads than calls from any other channel. However, once answered, these calls convert at about the average rate.
The report states that the lead rate for calls referenced by ChatGPT is 49%, about 10 percentage points higher than the average of the seven channels tracked by Invoca and 6 points above Google Business Profiles at 43%. The conversion rate for these leads is 40%, compared to an all-channel average of 42%. Invoca considers this to be average.
All figures represent Invoca customer averages, based on more than 70 million calls and 600 million talk minutes across 10 industries. Invoca sells the call tracking and conversation analytics that generate this data.
Invoca says this is the first year it has had enough data to measure calls generated by AI generative search.
What the data shows
Across all industries, approximately 56% of calls to businesses are answered personally. If a call lasts longer than 15 seconds, the answer rate increases to around 65%, and for calls longer than 30 seconds, it rises to around 71%. Of the calls answered, approximately 38% are considered leads, and approximately 42% of these leads are converted during the call.
ChatGPT is above this baseline for the first digit and below for the second.
Paid search continues to generate the most calls, leads, and conversions among paid channels in the dataset. For multi-location businesses, Google Business Profiles are the primary organic source. Invoca points out that channel efficiency and scale are different factors, and that percentages alone don’t reveal which channel generates the most business.
What the report doesn’t say
Invoca does not publish the number of calls referenced by ChatGPT from which the 49% is calculated, noting only that Invoca does not publish the number of calls referenced by ChatGPT from which the 49% is calculated, noting only that the overall volume attributable to generative AI remains very low. When a rate is derived from a small base, it tends to be less reliable than the same rate calculated from a significantly larger paid search volume.
The report does not specify a measurement window. The methodology explains that the figures are based on calls tracked and analyzed on the Invoca platform across 10 industries and seven marketing channels, but it does not mention a specific start or end date. Gemini, Claude and Perplexity are not included in the channel breakdown. Invoca notes that this is a measurement limitation rather than a comment on these assistants, mentioning that ChatGPT is the only large language model generating measurable call volume in its dataset.
How Invoca assigns calls
Invoca labels calls as referenced by ChatGPT, but the report lacks details on how this attribution works, such as whether callers clicked on ChatGPT, used tracked numbers, or contacted the company through other means. It only counts calls directly attributable to ChatGPT and not those from users who searched for a business in an assistant and then called via untracked methods.
I covered a version of this border back in June, when Similar web data linked ChatGPT brand recommendations to a 2.5x higher chance of visiting the site within seven days. Most related traffic appeared as a branded search rather than a direct referral, which limited what standard referral reports could display. The calls add another attribution problem because the report does not explain what digital trail Invoca used to connect them to ChatGPT.
Why it matters
Calls attributed to ChatGPT are considered leads more often than calls from other channels tracked by Invoca, by about 10 points. Once someone picks up, they convert at about the rate that companies can manage with everyone else. This complicates the reading that has formed around references to AI over the past year.
I wrote in May about Discovering Adobe that the conversion sign reversed on AI-referenced traffic to US retailers. In twelve months, it went from being the worst performing channel to having a 42% higher conversion rate than the others. The explanation offered was that the search had already taken place inside the assistant. Invoca’s data is the first half of that. Someone who compares options with an assistant and then calls may be further along in the purchasing decision, and that’s also how Invoca reads it.
The second half of the data doesn’t quite add up. Although a higher lead rate is observed, this does not translate into a higher on-call conversion rate when looking at Invoca’s averages. In this dataset, the difference appears at the qualification stage but then disappears.
Looking to the future
Invoca believes this is more of a signal to watch rather than a channel to invest in, supported by the volume caveat. The key metric that influences this view is the number of calls, which the report does not specify. Another question is whether the 40% are moving. If AI-referred callers continue to qualify at the top of the list while converting in the middle, the focus shifts from increasing call volume to understanding what’s happening during those calls.
The report also notes that 64% of businesses do not ask callers to make a purchase or schedule an appointment, which is a problem on the business side.
