ChatGPT Drives High-Quality Leads via Phone Calls but Struggles to Close the Deal

Calls referred by ChatGPT qualify as sales leads more often than any other tracked channel — but that early advantage disappears when it comes to actually converting those leads into customers.

A benchmark report published by Invoca on July 13, 2026, reveals a striking paradox at the intersection of AI-assisted research and real-world sales behavior. ChatGPT-referred phone calls lead all measured channels in lead qualification rates yet land squarely in the middle of the pack when it comes to converting those leads during the call itself.

For marketers and business leaders watching AI reshape the buyer journey, this data offers the clearest phone-call-level view yet of how generative AI influences purchasing behavior — and where its influence currently stops.

---

What the Numbers Actually Show

The Invoca report draws on more than 70 million calls and 600 million minutes of conversation across 10 industries and seven marketing channels. The dataset represents averages from Invoca's customer base, and the company notes this is the first year it had sufficient data to measure calls driven by generative AI search at all.

The headline figure is a 49% lead rate for ChatGPT-referred calls — approximately 10 percentage points above the all-channel average and 6 points above Google Business Profiles, which sit at 43%. In a world where most marketing channels compete for fractions of a percentage point, a 10-point lead rate advantage is a number that commands attention.

The conversion figure, however, tells a different story. ChatGPT-referred calls convert at 40% — just below the all-channel average of 42%. Invoca characterises this as approximately average performance, which means the quality signal that appears at the qualifying stage does not carry through to the close.

How ChatGPT Compares Across the Call Funnel

Across all industries, roughly 56% of calls to businesses are answered by a person. That rate climbs to about 65% for calls lasting more than 15 seconds and to approximately 71% for calls exceeding 30 seconds. Of answered calls, around 38% qualify as leads and roughly 42% of those leads convert during the call. ChatGPT sits above the baseline on lead qualification and just below it on conversion.

Paid search continues to generate the most calls, leads, and total conversions among paid channels in the dataset. For multi-location businesses, Google Business Profiles remain the leading organic source. Invoca is careful to distinguish between channel efficiency and channel scale, noting that percentage rates alone do not reveal which channel delivers the most business in absolute terms.

Putting the Data in Context

Understanding how AI-powered chatbots are transforming business operations and customer engagement is essential context for interpreting these numbers. ChatGPT is no longer simply a productivity tool — it is increasingly functioning as a pre-purchase research environment that shapes buyer intent before any direct brand contact is made.

---

The Attribution Gap That Complicates the Data

The report raises as many questions as it answers, particularly around how Invoca attributes calls to ChatGPT in the first place. The methodology does not specify whether callers clicked through from ChatGPT, reached the business via a tracked phone number, or made contact through some other traceable path.

This matters because attribution in an AI-assisted research environment is notoriously incomplete. A buyer might spend 20 minutes comparing options inside ChatGPT and then call a business directly by searching the company name or dialling from memory. That call would appear in analytics as branded search or direct traffic — not as a ChatGPT referral.

The Volume Problem Behind the Percentages

Invoca also does not publish the raw call count behind the 49% lead rate figure, noting only that total volume attributable to generative AI remains very low. A percentage derived from a small base is statistically less stable than one drawn from the much larger paid search volume. That caveat is significant when drawing channel-level conclusions.

Gemini, Claude, and Perplexity are absent from the channel breakdown. Invoca describes this as a measurement limitation rather than a judgment on those platforms, stating that ChatGPT is currently the only large language model generating measurable call volume in their dataset.

What Incomplete Attribution Means for Strategy

The absence of a standardised attribution model for AI-referred traffic is one of the most significant structural challenges facing marketers right now. Until platforms establish consistent tracking for generative AI referrals — similar to how UTM parameters function for traditional search — the true scale of AI's influence on inbound calls will remain underestimated. Marketers should treat current figures as a floor, not a ceiling.

---

Why the Lead Quality Advantage May Be Real — and Why It Stalls

The most compelling interpretation of the data centres on where buyers are in their decision process when they pick up the phone. Someone who has already compared products, reviewed options, and formed a preference inside an AI assistant may arrive at the call with fewer objections and more specific intent. That would explain why they qualify as leads at a higher rate.

The Pre-Call Research Effect

This framing aligns with findings reported elsewhere. Adobe data published earlier in 2026 found that AI-referred traffic to U.S. retailers had gone from the worst-converting channel to one that converted 42% better than other channels over the span of 12 months. The explanation offered at the time was consistent: the research happens inside the assistant before the buyer ever touches the brand's website or phone line.

Invoca's data supports the first half of that thesis. The qualifying stage reflects a more prepared buyer. The conversion stage, however, does not. Once the call begins, ChatGPT-referred callers behave much like everyone else — and the report offers one structural reason why. Invoca notes that 64% of businesses do not ask callers to make a purchase or schedule an appointment during the call. That is an execution gap on the business side and one that affects all channels equally.

Businesses investing in AI-driven customer service strategies and tools are better positioned to close this execution gap — particularly as inbound call volumes from AI-referred sources continue to grow.

The Conversion Ceiling and What Drives It

Invoca frames the current data as a signal worth monitoring rather than a channel ready for scaled investment. The volume remains too low and the attribution too uncertain to draw firm strategic conclusions. The more pressing question is whether the 40% conversion rate moves as AI search volume grows and as businesses improve how they handle inbound calls.

The data points to call handling as the most immediate lever within a business's control. A prepared buyer arriving via ChatGPT still encounters the same scripted responses, the same unanswered objections, and the same missed close attempts as any other caller — unless the business has specifically trained for high-intent inbound interactions.

Where Businesses Should Focus Now

For marketers and business operators, the report surfaces several implications worth acting on:

• Track AI-referred calls separately, even at modest volumes. Early data establishes the baseline against which future growth will be measured.
• Audit call handling for close attempts. The 64% of businesses that never ask for a commitment during the call are leaving conversions on the table across every channel — not just AI-referred ones.
• Optimise for the post-research buyer. As customer service chatbots and AI tools reshape the buyer experience, the businesses that win will be those prepared to meet a more informed, more decided caller — not one still in the early stages of consideration.

Invoca's data is the first credible measurement of AI's influence on inbound phone calls at scale. Its limitations are real, but its directional signal is clear: generative AI is already shaping who calls, and how ready they are to buy. The remaining gap — between a qualified lead and a closed sale — sits almost entirely within the business's own hands.

For a broader view of how AI referral traffic is evolving across digital channels, Invoca's full 2026 benchmark report provides additional industry-level breakdowns and methodology notes.