The news: OpenAI is laying the foundation to make AI a viable commerce channel, even before most consumers embrace generative AI shopping.

• The company launched Business Hub, giving businesses tools to manage their presence and product information in ChatGPT, and track insights such as how often they’re mentioned in ChatGPT responses and how many clicks they generate to their websites.
• It also rolled out Business Agents, which let consumers ask brands questions about their products. For example, someone shopping for a TV could ask a manufacturer to compare two models and recommend which would work better in a brightly lit room. Businesses can sponsor their agents and use them as destinations for ChatGPT ads, extending their reach to potential customers.
• OpenAI also pushed into travel with its acquisition of AI trip planning and booking platform Odessia.

Why this matters: AI platforms need a solid foundation before they can build a commerce ecosystem. OpenAI is putting the pieces in place by connecting business information, product feeds, branded interactions, and transactions. That infrastructure could allow commerce capabilities to grow alongside consumer adoption.

Early signs suggest AI is already shaping travel discovery and purchase decisions. AI referrals to US travel websites increased 119% YoY in July, according to Adobe, and their conversion rates have nearly caught up with those of non-AI traffic. Non-AI referrals converted just 1% better in July, compared with a 47% gap a year earlier. OpenAI’s new business tools could help travel brands capitalize on that momentum by improving their visibility in AI recommendations and engaging consumers earlier in the decision-making process.

But several roadblocks stand in the way of widespread adoption. First, AI platforms need access to the full range of options available to consumers when they shop independently, and for now, they don’t have it. Amazon is blocking bots from platforms such as ChatGPT and Muse from accessing its site, while United Airlines told The Wall Street Journal on Wednesday that it is preventing bots from accessing its systems to book tickets.

Second, AI platforms need to give consumers a clear reason to change their habits by solving a problem they actually have, or at least one they recognize. It’s not yet clear that general-purpose AI agents do that, particularly when it comes to travel. Airbnb CEO Brian Chesky, for instance, has questioned whether consumers will rely on a single agent for everything. He said agents such as Muse and Instinct are suited to simple, routine tasks like making restaurant reservations or scheduling appointments but don't necessarily add value to travel, where discovery is an intrinsic part of the experience.

Chesky noted that 80% of travelers enjoy researching trips: browsing photos, comparing properties, and making decisions with travel companions. That may help explain why only a negligible share of Airbnb’s roughly 500,000 daily bookings originate from Muse and Instinct, he said.

Implications for marketers: We expect US generative AI users to rise 47.6% this year to 161.5 million, with growth accelerating every year through 2030. That large user base will give these platforms plenty of opportunities to test, learn, and iterate on how agentic commerce can and should work.