支付、互联网与零售：智能代理时代——首届智能代理商业日洞察

The Age of Agents: Insights from our Inaugural AgenticCommerce Day We hosted our inaugural Agentic Day in New York last week. This note contains our cross-sector takeaways from the event. Harshita Rawat, CFA+1 917 344 8485harshita.rawat@bernsteinsg.com Speakers included Fiona Tan, CTO at Wayfair, Sherri Haymond, EVP for Global DigitalCommercialization at Mastercard, Max Zevin, Managing Director & Partner at BCG, CaseyCarl, Former EVP and Chief Omni Strategy and eCommerce at Walmart, Chief MarketingOfficer at GLD, Chris Wichert, Founder and former CEO of Koio, Matthew Eberhart, andRichard Btaiche, Product Lead at Paxos. Key Takeaways: Nikhil Devnani, CFA+1 917 344 8425nikhil.devnani@bernsteinsg.com Aneesha Sherman+1 917 344 8457aneesha.sherman@bernsteinsg.com Agentic commerce is very much in the iterative phase with product market fit -and not ready to take over the full (and often asynchronous) shopping journeys yet.Consumers are increasingly using AI tools for discovery, but not yet transacting – they havemaintained relationships with their favorite marketplaces and retailers. Under 1% of webtraffic can be attributed to referrals from AI chatbots today for the largest players, thoughit’s as high as 10% for the DTC brands we hosted.The grand vision around Agentsbuying on our behalf is increasingly looking a bit far out, especially for medium tohigh consideration categories. Instead, we had a lot of discussions about AI-enhanced/conversational commerce experiences and improved product discovery both on leading AIapps (e.g., ChatGPT and Gemini) as well as on merchants’ owned and operated properties(e.g., Wayfair building out an AI-assisted shopping experience). Zhihan Ma, CFA+1 917 344 8303zhihan.ma@bernsteinsg.com Mark Shmulik+1 917 344 8508mark.shmulik@bernsteinsg.com Viola Chen+1 917 344 8614viola.chen@bernsteinsg.com OpenAI’s evolving playbook(including recent abandonment of Instant Checkout)highlights the difficulty of changing consumer behavior, the importance ofintegrations/underlying infrastructure, and challenges with checkout(e.g., onconversion, real-time pricing, post-purchase, syncing of merchant catalogs & inventory,fulfillment and multi-basket). Google’s shopping graph gives it a better starting point in thatregard. Automation could be the real unlock in the customer value proposition that drivesbehavior to change.However, as the conversation has tilted towards ‘discovery’, thefuturistic scenario of delegated Agent-to-Merchant transactionsfeels further out fornow. Simran Ratani+1 917 344 8329simran.ratani@bernsteinsg.com Nathan Gee+1 917 344 8573nathan.gee@bernsteinsg.com Jessica Tian+1 917 344 8413jessica.tian@bernsteinsg.com Merchants, however, are preparing — the big challenges are tech stack and Jed Hodulik+1 917 344 8594jed.hodulik@bernsteinsg.com payments. BCG noted that merchants are really taking note – 70% of the merchants theysurveyed are already working on Agentic Commerce discovery. They expect that to rise to90%. For merchants, it’s critical that catalogs are well maintained, current, and feed intothe shopping experiences on ChatGPT/Gemini. Much of the same traits that drive SEOappear to be determining GEO, and the leading retailers are also using AI tools to furtherenhance and optimize their listings (e.g., better imaging, better product descriptions, etc.).Integrations with promotions and rewards programs are needed to scale these offeringssuccessfully. Additional areas where AI is driving ROI already is on marketing efficiency andcustomer support. On the payments side, a lot of discussion around risk and liabilities if anAgent is making recommendations or buying on your behalf – it’s a huge challenge. Jeremy Miles, CFA+1 917 344 8370jeremy.miles@bernsteinsg.com Wenhuan Chang+1 917 344 8546wenhuan.chang@bernsteinsg.com Deeksha Pandey+1 917 344 8447deeksha.pandey@bernsteinsg.com Continued on next page... Questions raised about scale and brand value.Large enterprises and leadingmarketplaces have the capacity and talent to build/integrate directly with the leadingAI labs (e.g., Wayfair partnered with Gemini to launch UCP, Amazon launched Rufus).Elsewhere, Shopify is bringing along the long tail. It’s really mid-market that is fallingbehind. There is also debate about the value of a brand in an AI world, considering brandsare often shorthand determinations for trust building, and AI tools can surface a broaderarray of merchandise across more vectors. To that end, DTC brands stand to benefit inthe short-term as AI makes their products more discoverable, particularly within theirrespective niches. Longer-term however, we’d expect larger brands to catch up, andsuperior fulfillment networks and pricing power give them an edge when it comes to GEO,traffic generation, and building superior AI-enhanced shopping experiences. Horizontal vs. Vertical AI experiences.As the leading AI labs tinker with eCommerceexperiences, the leading marketplaces are launching their own off