EMARKETER data: consumers rarely use AI chatbots to shop
New EMARKETER research finds that AI chatbots drive product discovery but rarely convert into completed purchases.
Mariano De Vitto · August 2026
New research from EMARKETER, cited via Insider Intelligence, delivers a reality check to marketers betting big on conversational commerce: consumers are still overwhelmingly reluctant to complete purchases inside AI chatbots like ChatGPT, Gemini or Meta AI. Despite months of hype around chat based shopping assistants, the data shows a persistent gap between platform ambition and actual consumer behavior at the point of purchase.
Over the past year, OpenAI, Google and Meta have all pushed to turn their conversational AI products into commercial destinations, embedding product recommendations, links and even checkout style flows directly into chat interfaces. Brands, eager to reach consumers wherever attention is shifting, have followed by testing ads and sponsored placements inside these tools. But EMARKETER's findings suggest that enthusiasm from the platform and advertiser side has outpaced actual consumer readiness to buy.
Research Yes, Checkout No
According to the report, consumers are increasingly comfortable turning to AI chatbots for research, comparison shopping and general product questions. Where they hesitate is at the final step: handing over payment details or completing an order within the same conversation. That reluctance mirrors a pattern long familiar in digital retail, where discovery and research often happen on one platform while the actual transaction closes somewhere else, be it a retailer's own site, a marketplace like Amazon or a physical store.
This split between research and transaction is not new to marketers who have watched social platforms, search engines and even voice assistants go through similar cycles. Consumers build trust in a new interface gradually, using it first for lower stakes tasks like comparing prices or reading summaries before they are willing to expose sensitive information like credit card numbers. AI chat is simply the latest channel following that familiar arc.
What This Means For Budgets
For marketers who have been eyeing chatbot placements as the next major advertising inventory, the message from EMARKETER is clear. Top of funnel value, meaning increased product awareness and consideration, appears to be real and growing. Bottom of funnel payoff, meaning direct sales attributable to a chatbot conversation, is not yet proven at scale. Treating these placements as if they were a mature direct response channel risks disappointing return on investment.
Media buyers currently testing budgets against OpenAI, Google and Meta's conversational products should reframe expectations accordingly. These placements are best positioned today as awareness and consideration plays, similar in spirit to early display or social advertising, rather than as performance channels expected to drive immediate transactions. Measurement frameworks should reflect that reality, tracking brand lift, engagement and assisted conversions rather than last click sales from chat.
Next Steps For Marketers
Why it matters: brands should temper their expectations for AI chatbots as a direct sales channel for now, and instead prioritize them for discovery, product education and brand visibility. Budget allocation should favor small scale testing and learning over heavy upfront investment, giving teams room to understand how consumers actually behave inside these interfaces before committing significant spend.
As EMARKETER and other research firms continue to track this space, marketers should watch closely for signals that purchase behavior inside chatbots is shifting. Until that evidence arrives, the smartest move is to use AI chat for what it already does well: helping consumers discover and evaluate products, while leaving the closing of the sale to the channels where consumers still feel comfortable completing it.
The Signal Brief · Mariano De Vitto — Head of Marketing, Barcelona