limitedDistribution · Industry Research
Beauty in Agentic Commerce
Agentic commerce is a retail model in which AI agents help shoppers discover, compare, and buy products inside one conversational flow. According to Paz.ai and.
Agentic commerce is a retail model in which AI agents help shoppers discover, compare, and buy products inside one conversational flow. According to Paz.ai and PYMNTS.com, the model connects the consumer, the AI agent, and the merchant through standardized protocols that support discovery, evaluation, and purchase without forcing the shopper through separate search and checkout steps. In practice, the agent evaluates product attributes such as price, availability, shipping speed, reviews, and relevance to the shopper’s stated intent. That changes how merchants compete: catalogs that are complete and machine-readable are more likely to be recommended, while incomplete or unstructured product data can be skipped. PYMNTS.com also reports that Ulta Beauty is adding shopping carts and its loyalty program into Google’s Gemini because customers are already finding Ulta products there, though the effort remains early.
Key Takeaways
- Agentic commerce matters now because the market has crossed from experimentation into measurable retail impact.
- Trend 1: AI shopping is becoming mainstream research behavior before it becomes mainstream checkout behavior.
- Trend 2: AI shopping surfaces are moving from referral engines toward transaction channels.
- Trend 3: Speed is becoming the strategic moat, from product development to personalization.
- For retail operators, agentic commerce turns catalog quality, feed operations, and measurement into day-to-day revenue infrastructure.
Agentic commerce matters now because the market has crossed from experimentation into measurable retail impact. According to PYMNTS.com, AI-referred traffic to U.S. retail sites grew 393% year over year in the first quarter of 2026, and revenue per visit from those referrals was 37% higher than non-AI traffic as of March 2026. That indicates AI agents are not only influencing discovery but also sending higher-value shoppers to merchants. The infrastructure is also catching up. Paz.ai says consumer adoption, production-ready platform infrastructure, and emerging open protocol standards have made agentic commerce operational. The shift has happened quickly: agentic commerce moved from research demos to live consumer transactions in roughly 18 months, driven by protocol launches from Anthropic, Stripe, OpenAI, and Google. Consumer-facing proof points are now visible as well, with ChatGPT Shopping launching with Etsy, Glossier, SKIMS, Spanx, and Vuori in September 2025. Together, these signals suggest agentic commerce is becoming a near-term operating channel rather than a future concept. The clearest near-term shift in agentic commerce is not that consumers are handing over the entire purchase to an AI agent. It is that they are already using AI at the top and middle of the buying journey: discovering options, comparing products, narrowing choices, and preparing to buy. PYMNTS.com reports that 48% of online shoppers used AI to research their most recent purchase, while also summarizing the current retail pattern as consumers allowing AI to find products but fewer allowing AI to complete purchases. That distinction matters for merchants. AI-assisted shopping is already large enough to influence traffic quality, product consideration, and conversion paths, even if autonomous checkout remains less common. Paz.ai cites OpenAI data from February 2026 showing ChatGPT processes 50 million shopping queries daily, and also cites a January 2026 IBM Institute for Business Value study finding that 45% of consumers already use AI for some part of the buying journey. The same direction appears in consumer adoption data: Paz.ai cites a Morgan Stanley AlphaWise survey of 2,000 U.S. consumers in November 2025 showing 45% had used ChatGPT in the prior month, and Adobe’s 2025 finding that 38% of consumers had used generative AI for online shopping. For retailers, the implication is practical: optimize for being found, understood, and recommended by AI systems now, while payment trust and delegated purchasing catch up. At the same time, AI shopping surfaces are moving from referral engines toward transaction channels. AI assistants are no longer just answering product questions; they are increasingly shaping where discovery, comparison, and checkout begin. Paz.ai reports that ChatGPT Shopping is live for all U.S. users, recommends products, and directs shoppers to merchant sites. The same report says Etsy and more than 1 million Shopify merchants are live on ChatGPT Shopping, showing that AI shopping is already connected to large merchant ecosystems rather than remaining a niche experiment. This is also becoming a platform race. Paz.ai notes that Google launched AI Mode with agentic checkout capabilities, with Wayfair, Chewy, and Etsy among early participants. It also reports that Microsoft Copilot supports checkout via ACP, while Perplexity offers instant purchasing through a PayPal-powered shopping feature. Together, these moves indicate that discovery, cart creation, and payment are starting to collapse into a single assistant-led flow. Demand signals are following the infrastructure. According to PYMNTS.com, Shopify data shows AI-referred traffic to Shopify stores tripled year over year in the second quarter of 2026, and orders through Shopify AI channels also tripled over the same period. PYMNTS.com also cites Shopify President Harley Finkelstein as saying new buyer orders from Shopify AI channels are arriving at nearly twice the rate of other channels. For merchants, that makes agentic commerce less about future speculation and more about preparing product data, offers, and checkout paths for AI-driven buyers now. For beauty in particular, speed is becoming the strategic moat, from product development to personalization. The next competitive divide is not only who has the strongest brand, but who can learn, formulate, launch, and personalize fastest. According to Nikkei Asia, South Korea’s cosmetics industry has become the world’s second-largest exporter of beauty products, with that rise tied to contract manufacturers that help bring new products to market quickly. The key operating model is a sharp division of labor: ODM companies such as Kolmar do not carry their own brands and instead focus on R&D and contract manufacturing, while brand owners can concentrate on consumer insight, positioning, and demand creation. Nikkei Asia reports that this structure helps the industry produce hit products at unparalleled speed. That same speed logic is now extending into digital retail. PYMNTS.com reports that Etsy has built richer buyer profiles covering more than 65 million shoppers to support real-time personalization in search results, home feed content, and marketing outreach. In practice, this points to a broader shift: beauty and retail players are trying to shorten every feedback loop, from detecting demand signals to presenting relevant products to individual shoppers. For operators, the implication is clear. Fast manufacturing capacity alone is not enough, and personalization alone is not enough. The advantage comes when product development systems, supplier networks, and customer data infrastructure reinforce each other, allowing companies to respond to trend shifts with both relevant products and relevant experiences.
Operational Impact
For retail operators, agentic commerce turns catalog quality, feed operations, and measurement into day-to-day revenue infrastructure. According to @paz.ai, enterprise retailers should audit product data, enrich catalogs for natural-language intent, publish channel-specific feeds, select commerce and payment protocols, and measure whether AI agents recommend their products. That shifts work across merchandising, product information management, ecommerce, payments, and analytics teams rather than leaving agent visibility as a search or SEO task alone. The first operational impact is data readiness. Retailers need to score representative product pages for structured attributes, product context, availability, and schema gaps before deciding how to integrate with agentic channels. They also need AI catalog management to close missing-data gaps and AI merchandising practices that make attributes and descriptions match real shopper questions. The second impact is feed and live-data discipline. Teams must map required fields, channel formats, inventory updates, and pricing refreshes so agents receive current, usable product information. Weakness here can create stale product decisions or product invisibility, two risks @paz.ai identifies for agentic commerce. The third impact is governance and performance monitoring. Retailers should establish an AI visibility baseline and track found rate, position, competitor presence, and changes after catalog releases. They also need to evaluate transaction authorization and unit economics, because @paz.ai identifies unsafe authorization and weak unit economics as additional risks.
What Buyers Should Evaluate
- Buyers evaluating agentic commerce capabilities should separate customer-facing discovery from payment execution and data connectivity. According to @paz.ai, the Agentic Commerce Protocol supports discovery and commerce flows, while AP2 addresses scoped, verifiable payment authorization. That distinction matters because a solution that improves product discovery may not automatically provide the authorization controls needed for autonomous checkout. Buyers should also assess which protocol ecosystems a vendor supports. @paz.ai notes that the Agentic Commerce Protocol was developed by Stripe and OpenAI and powers ChatGPT Shopping product discovery and merchant redirect, while Google’s Universal Commerce Protocol covers the commerce journey from discovery to post-purchase, and Anthropic’s Model Context Protocol provides agent access to real-time inventory, pricing, and product details. In practice, evaluation should cover whether the platform can expose accurate product data, route customers to the right merchant experience, and support future post-purchase workflows. Demand readiness is another filter. PYMNTS.com reports that 49% of consumers would let an AI agent complete a purchase, but trust drops to 3% among consumers who do not already use AI regularly. Buyers should therefore look for clear consent flows, transparent authorization boundaries, and experiences that build confidence for first-time AI commerce users rather than assuming broad adoption will happen immediately.
Definitions
Agentic commerce: According to @paz.ai, agentic commerce typically unfolds in four stages: intent capture, discovery and matching, evaluation and recommendation, and transaction execution. Agentic Commerce Optimization: @paz.ai defines Agentic Commerce Optimization as the discipline of optimizing merchant catalogs for AI-agent discovery and matching. Agentic Commerce Protocol: @paz.ai says the Agentic Commerce Protocol was developed by Stripe and OpenAI and powers ChatGPT Shopping product discovery and merchant redirect. Universal Commerce Protocol: @paz.ai reports that Google’s Universal Commerce Protocol was launched at NRF 2026 and covers the commerce journey from discovery to post-purchase. Model Context Protocol: @paz.ai describes Anthropic’s Model Context Protocol as a data connectivity layer that lets agents access real-time inventory, pricing, and product details. ODM companies: Nikkei Asia reports that ODM companies such as Kolmar do not carry their own brands and focus only on R&D and contract manufacturing.
FAQ
Q: What is agentic commerce? A: Agentic commerce refers to shopping journeys where AI agents or assistants help discover, compare, recommend, and potentially initiate purchases on behalf of consumers. The current opportunity is large: According to @paz.ai, McKinsey projects agentic commerce will generate $3 trillion to $5 trillion annually by 2030. Q: Is this already a major traffic channel for retailers? A: Not yet in every category. PYMNTS.com reports that Etsy CEO Kruti Patel Goyal said traffic from AI agent platforms remained under 1% of Etsy’s total traffic, though that traffic converted at a higher rate and produced a larger average order size. That suggests the channel is still early, but the visitors it does send can be commercially valuable. Q: How do AI-referred shoppers behave differently? A: PYMNTS.com, citing Adobe market data, reports that shoppers arriving from AI assistants spent 48% more time on the page and browsed 13% more pages per visit than traffic from other channels. @paz.ai also cites Adobe Analytics data from Black Friday 2025 showing AI-referred visitors had a 38% higher purchase completion rate than traditional search visitors. Q: What is the biggest operational gap? A: The main unresolved issue is the handoff between AI shopping assistants and fulfillment. PYMNTS.com reported that this handoff remains unsolved industry-wide, meaning retailers still need reliable processes for inventory accuracy, order confirmation, payment, delivery, returns, and customer support when an AI agent is involved. Q: How big could the market become by 2030? A: Estimates vary, but they point in the same direction. @paz.ai cites Morgan Stanley AlphaWise estimating $190 billion in U.S. agentic commerce by 2030, roughly 10% of U.S. e-commerce, while Bain & Company estimates 15% to 25% of U.S. e-commerce by 2030.
Stargo Insight: Agentic Beauty Commerce Will Stress Post-Purchase Operations
Agentic commerce is often framed as a discovery and checkout shift, but beauty retailers should watch the operational tail: returns documentation, refund decisions, and exception routing. Stargo e-commerce benchmarks show AI-assisted returns document validation reduced manual case routing by 33% during peak season operations, while merchants using AI triage scoring routed high-risk documentation exceptions 2.3x faster. For beauty brands selling through AI-led channels, that post-purchase speed can become as important as catalog readiness, especially when shoppers expect fast resolution on damaged, mismatched, or policy-sensitive orders.
Related guides: Investment Platform Trends for E-commerce Buyers, Digital Transformation in E-commerce.
See ROI in 12 weeks
Stargo users see measurable return and operational profitability gains in just 12 weeks, with non-disruptive implementation in 4 weeks or less.