PayPal’s play for agentic commerce: Making AI shopping work for Main Street
- PayPal is building infrastructure to connect merchants to AI shopping platforms like ChatGPT and Gemini through a single integration, positioning itself as the layer between commerce and conversational AI.
- The company argues this could level the playing field for small merchants by shifting discovery from ad-driven search to AI recommendations based on product fit.
AI-powered shopping is moving faster than most merchants can keep up with. ChatGPT, Perplexity, and Google’s Gemini are all racing to let consumers buy products directly through conversational interfaces, but the infrastructure to make that actually work remains fragmented. AI platforms want their own integrations, the protocols all look different, and merchants are left trying to figure out which standards to adopt while their core business demands attention elsewhere.
PayPal’s positioning itself as the infrastructure layer that connects all of it.
I sat down with Michelle Gill, executive vice president and general manager of PayPal’s Small Business and Financial Services Group, at Money20/20 to talk about the company’s move into agentic commerce. The pitch: merchants are drowning in new protocols — MCP servers, API variations, different LLM integrations — and they don’t have time to figure out which horse to bet on. Meanwhile, PayPal expects over 40% of Americans will start product discovery on AI surfaces this holiday season. That’s discovery, not purchases, but the behavior shift is already happening.
“Every day there’s a new protocol announced,” Gill told me. “Merchants are trying to figure out how much time to spend being agentic-friendly, and it’s already a stretch because they have no time for anything.”
PayPal’s solution: become the abstraction layer between merchants and every AI shopping surface. One integration that propagates to ChatGPT, Gemini, Perplexity, and whatever comes next. Orders flow back through existing infrastructure instead of requiring separate reconciliation systems for each platform.
OpenAI’s recent announcement, where PayPal’s wallet gets baked into ChatGPT’s shopping experience, shows the strategy in action. PayPal worked with OpenAI to create an open wallet protocol. They’re positioning themselves as neutral infrastructure that makes the whole thing work, not as a destination competing with these platforms.
The SMB angle is where this gets interesting. Gill’s argument is that without centralized infrastructure, large retailers will dominate early AI shopping integrations while small merchants get left behind. The gap that already exists in digital advertising — where whoever spends most on ads wins — could get worse in AI commerce. But if discoverability shifts from search ranking to conversational AI, the economics change. Small merchants wouldn’t need to outspend competitors on AdWords to get discovered. An AI agent surfaces products based on fit, not ad budget.
“If discoverability is no longer based on beating up AdWords, you actually have potentially neutralized that disparity,” Gill said.
Whether that plays out depends on how AI platforms structure their recommendation logic. If it’s still pay-to-play, nothing changes. If it’s genuinely based on matching consumer needs to product fit, PayPal’s infrastructure bet makes more sense. The company doesn’t need to own the shopping experience; they need to enable it everywhere it happens.
The technical challenge is order management: confirming inventory exists, processing the transaction, calculating taxes and shipping across surfaces where PayPal doesn’t control the UI. They’re building merchant onboarding tools now, starting with demos for early partners. Gill calls this a strategic priority at PayPal, ranking alongside core products. The company is staffing it with cross-functional teams repurposing existing architecture for new use cases. They’re also working with Nvidia on reducing compute requirements, since general-purpose models burn expensive processing power for relatively simple shopping tasks.
The consumer trust piece is where PayPal thinks it has an edge. Gill draws parallels to the original eBay days: consumers didn’t want to give credit card details to random sellers in other states, so PayPal became the abstraction layer. Now it’s: consumers might not want their full payment credentials stored in an AI agent, so PayPal becomes that layer again.
Security is the biggest risk to adoption, according to Gill. If trust breaks early, the whole thing stalls. The near-term reality is more mundane: merchants need conversational search interfaces that match consumer expectations set by large retailers. PayPal wants to give them that capability without requiring engineering resources many of these merchants don’t have.
PayPal stock jumped 9% when it made its agentic announcement. Whether that holds depends on merchant adoption, which depends on consumer behavior shifting faster than it has so far. But with 400 million consumers and 20 million merchants already in the network — and 2 billion more through international wallet integrations — the distribution is there. The question is whether being infrastructure instead of a platform owner matters as much as PayPal thinks it does.