Agents, digital wallets, and trust are rewriting checkout

Stripe analyzed payment data from roughly 20,000 B2C businesses covering August 2023 to February 2026, finding that checkout is no longer just a simple payment step. Mobile, region-specific payment methods, and AI agents are becoming the key variables that determine payment outcomes.

Mobile payments are spreading beyond small purchases into high-value purchases as well. On Stripe, 65% of transactions under $50 occur on mobile, and in the US, mobile's share has grown across all purchase amount ranges over the past two years. Canada is a notable exception, however, showing a shift toward desktop in the $100–$249 range.

Digital wallets are establishing themselves as the base layer of global checkout. About 30% of global POS transactions are made via wallets, and in a Stripe survey, 61% of consumers expressed willingness to use digital wallets. In particular, respondents aged 18–29 said they use wallets for 50% of purchases under $25 and 33% of purchases over $250, with speed—cutting mobile checkout time in half—being the key driver.

However, even as wallet adoption grows, preferences vary widely by market. Because the top payment method differs by region—MB WAY in Portugal, MobilePay in Denmark, Alipay in China—simply adding more payment methods isn't enough. In Stripe's experiments, showing even a single geographically mismatched payment method reduced conversion by up to 15%, while offering BLIK in Poland raised conversion by an average of 46%, and offering Pix in Brazil raised it by 31%.

As global demand grows, checkout needs to become more market-tailored. In markets with fragmented payment preferences like Indonesia and Vietnam, businesses need to localize the mix of payment methods, currency, and how options are displayed, whereas in markets with concentrated preferences, they should focus on a single dominant method.

The final shift is AI agents. In a survey of over 3,500 consumers conducted by Stripe and Visa, a majority in multiple markets said they would be willing to accept purchase assistance from AI agents. Meanwhile, as automated attacks like card testing increase, businesses face a dilemma where raising risk thresholds can also block legitimate customers—something that more sophisticated real-time signal evaluation, selective authentication, and routing/retry optimization can help mitigate. Stripe states that such AI-driven interventions achieved a 30% reduction in fraud without lowering conversion.

Ultimately, checkout is transforming into the point where user identity, purchase intent, and approval authority are verified. Product discovery and purchasing have already begun happening within AI interfaces such as Google Gemini, Microsoft Copilot, OpenAI's image-based shopping, Stitch Fix Vision, and Walmart Sparky, and companies now need to design payment flows that can reliably verify both humans and agents.