The Maze: The regional scorecard turns Morgan Stanley's Five I's into a useful warning. Wayfair leads the selected field at 23/25. Walmart follows at 22. Amazon, Mercado Libre, Alibaba and Shopee score 21. Yet the leaders win for different reasons—and the gap between technical readiness and economic value is hiding inside the component scores.
Wayfair wins on balance, not scale. It scores 5 on Inventory, Infrastructure, Innovation and Margin, with Growth the only 3. Walmart and Amazon are almost perfect across the first four tests, but both score only 2 on Margin. That is the core agentic-commerce paradox: the largest traffic owners can be easiest for agents to use and most exposed when free discovery, retail media or direct demand becomes a paid referral. Morgan Stanley's original report makes the same directional case, highlighting Wayfair's technology stack, early AI lead and CastleGate logistics while noting its lower retail-media downside. The operating evidence has since strengthened: Wayfair co-developed Google's Universal Commerce Protocol and will remain merchant of record for pricing, fulfilment and support in eligible agent-led checkouts.
Mercado Libre is Latam's integrated outlier. Its 21 is five points above Magazine Luiza and eight above Casas Bahia. The score is built on 5s in Infrastructure and Growth rather than a perfect assortment or margin profile. The company's 2025 filing helps explain why: fulfilment centers handle more than half of shipments; Mercado Envios connects warehouses, cross-docking, partner stores and last-mile carriers; and the network now ships almost all items sold. Mercado Libre reported 121 million active buyers, $65.0 billion GMV and $1.327 billion of capital expenditure in 2025. Agents like clean data. They love dependable physical execution.
APAC has the deepest selected bench. Alibaba and Shopee score 21; Meituan, Coupang and Nykaa score 20. All five receive a 5 on Inventory, and four receive a 5 on Infrastructure. No other displayed region packs five companies into a two-point band at the top. Shopee's disclosures show the operating machinery behind the score: more than $100 billion GMV and 10 billion orders in 2024, almost half of Asian SPX Express orders delivered within two days in Q4, and AI tools improving search, listings, customer service and returns. Sea said AI reduced customer-service cost per contact by nearly 30% and improved return-resolution times by more than 40% in Asia. That is not an AI demo. It is an operating system.
EMEA's constraint is supply and growth, not a lack of experiments. Allegro and Kaspi tie at 15; Zalando scores 14. All three receive only 2 on Inventory, while Innovation lands at 3 or 4. Allegro's Q1 2026 investor material describes a live Shopping Assistant, AI catalogue checks, improved intent matching and explainable product recommendations. The lesson is uncomfortable but useful: an excellent agent interface cannot compensate indefinitely for a thinner assortment or weaker incremental-growth opportunity. Conversation is not selection. Selection is not scale.
The weakest layer is the one that pays the bills. Across the 17 selected rows, Infrastructure averages 4.12/5—the strongest component. Margin averages 3.41, narrowly below Growth at 3.47. The technology is advancing faster than the proof of additive economics. Morgan Stanley's sensitivity work suggests roughly half of agentic transactions may need to be incremental for EBIT break-even at a 5% fee. Its public market outlook puts US agentic-commerce spending at $190 billion to $385 billion by 2030, or 10% to 20% of ecommerce. That can be a large channel shift without being free growth.
Why it matters: Retailers should read the Five I's as two gates. Inventory, Infrastructure and Innovation determine whether an agent can see, trust and fulfil the offer. Growth and Margin determine whether the resulting order creates enterprise value. Build structured catalogues, accurate availability, delivery promises and agent-ready checkout. Then measure new-customer rate, displaced direct traffic, retail-media leakage, return cost and contribution margin by agent. The winners will not be the retailers with the most AI. They will be the retailers that convert machine readability into profitable incrementality.
Method note: Exact retailer scores come from the supplied LinkedIn exhibit. Region sizes differ, and EMEA contains only the three companies shown in that source; the table is a selected scorecard, not a global census.
