The Maze: ChatGPT is becoming less of a blank page and more of a shopping aisle. Profound estimates that commercial-intent conversations rose from 13.9% to 19.2% of activity between June 2025 and June 2026. Informational use barely moved, while generative use lost 6.4 percentage points. The important shift is not that people stopped asking ChatGPT to write. It is that a much larger share now asks what to buy, which product fits, and where to get it. Commerce is moving upstream—from the product page into the conversation that defines the shortlist.
Commercial intent gained 5.3 percentage points in one year. The Profound analysis classifies 7.5 million conversations across 52 weeks. Commercial prompts increased from 13.9% to 19.2%, a 38% relative rise. Informational prompts edged from 40.1% to 41.2%. Generative prompts fell from 46.0% to 39.6%. The composition shift matters more than the headline growth: almost one in five sampled conversations now carries commercial intent, even though no checkout needs to happen inside ChatGPT for a brand to win or lose consideration.
The 533 million weekly estimate is an extrapolation, not a platform count. Profound combines 0.54 commercial conversations per user per week with an assumption of up to one billion weekly users. That produces the 533 million estimate. OpenAI had officially reported more than 900 million weekly active users earlier in 2026, so the order of magnitude is plausible. But Profound does not disclose enough about Prompt Volumes sampling to show that its prompt mix represents all ChatGPT users. The direction is useful. The absolute volume deserves a confidence interval, not a victory lap.
The commercial value appears before the click. A shopper can ask for a comparison, reject two options, refine a constraint and leave with one brand in mind—without generating a referral session. That makes conventional last-click reporting especially weak. The original discussion notes that commercial conversations run longer than informational or generative ones, which suggests more opportunities for brands, merchants and ads to enter the decision. It also creates a measurement problem: zero-click influence can grow while publisher and retailer analytics see less of the journey.
OpenAI is building the supply side for this behavior. Its product-discovery program gives merchants a route to submit structured product data so recommendations can use fresher availability, price and catalog information. Its shopping research experience is designed to ask clarifying questions, compare products and produce a buyer's guide. Those products do not validate Profound's percentages. They do validate the strategic direction: OpenAI expects shopping discovery to become a native conversational job.
Brands need an influence ledger, not another traffic dashboard. The operating response is straightforward. Keep feeds complete and machine-readable. Track whether products appear for high-intent questions. Compare recommendation visibility with branded search, direct traffic and conversion lift. Run controlled tests around catalog changes and answer coverage. The goal is not to attribute every sale to one invisible prompt. It is to learn whether conversational discovery changes demand before the shopper reaches a channel that can be measured.
Why it matters: Ecommerce has spent two decades optimizing the page after the click. Conversational discovery shifts leverage to the answer before the click. If commercial intent keeps gaining share, product data, recommendation visibility and brand credibility become distribution infrastructure. Profound's sample cannot tell us the exact size of that market. It can tell operators where to start looking—and why waiting for referral traffic to prove the shift will leave them late.
