The Shift Toward Autonomous Shopping
Venture capitalists are pouring capital into AI agents designed to handle transactions on behalf of consumers. These tools aim to bridge the gap between simple chatbots and autonomous digital assistants capable of executing purchases with a credit card. While the technology promises to change how people interact with e-commerce, the industry faces significant hurdles regarding trust, security, and technical reliability.
Recent data reflects a shift in consumer habits. An Adobe Analytics survey indicates that 41% of respondents used AI for online shopping as of June. Furthermore, a report from Retail Dive and Rithum suggests that 53% of shoppers now view AI recommendations with the same level of confidence as brand websites. Investors like Mark Grace of M13 observe that consumers are increasingly comfortable offloading routine tasks to software. Still, others remain cautious, noting that shopping is often a leisure activity rather than a chore people want to outsource to machines.
Categories and Current Limitations
Startups in this space generally fall into five distinct buckets. Personal assistants, such as Instinct and Town, attempt to handle tasks ranging from dinner reservations to travel bookings. Despite their ambition, these platforms remain prone to errors. One user recently reported a $300 loss after an agent canceled a flight without proper notice. Such incidents highlight why broad consumer adoption remains distant; people are hesitant to grant financial control to tools that make costly mistakes.
Other companies focus on discovery rather than direct purchasing. Daydream, founded by Julie Bornstein, provides personalized AI search tools that embed directly into retailer websites. This approach allows brands to maintain control over the customer experience while offering AI-driven navigation. However, the sector also faces reputational risks. The shopping assistant Phia recently faced allegations of cookie-stuffing—a practice where a company claims credit for sales it did not drive. While the startup disputes these claims, the controversy underscores the lack of industry-wide standards for attribution.
Infrastructure and the Path Forward
Technological development is moving beyond simple front-end interfaces. Companies like Catena Labs and Basis Theory are building the back-end infrastructure needed to verify agent identities and facilitate secure payments. Meanwhile, marketing tools such as Profound are pioneering Generative Engine Optimization to ensure brands remain visible in an AI-driven search environment. These companies have attracted significant venture capital, with Profound securing $96 million in Series C funding earlier this year.
Major players are also formalizing the rules of engagement. OpenAI and Stripe previously introduced the Agentic Commerce Protocol to standardize how agents interact with merchants. Google, Visa, and Mastercard have launched similar frameworks to integrate AI into existing financial systems. Success remains elusive for some initiatives; OpenAI pulled its Instant Checkout feature in March to reevaluate its product discovery strategy. As these protocols mature, the industry must solve the underlying tension between convenience and the inherent risks of automated finance. For now, agentic commerce occupies a small fraction of global trade, but the structural foundations are being laid behind the scenes.