AI is resetting commerce. In B2B, the stakes are ten times bigger.

The new McKinsey report outlines five shifts reshaping European e-commerce. In enterprise B2B, each of those shifts has a version that's more complex, more structural, and more defensible. Who's already building it?

In June 2026, McKinsey published Europe's new e-commerce agenda: How AI is resetting growth and competition — a sharp analysis of how artificial intelligence is redefining the economics of European digital commerce. Five structural shifts, four value levers that form a flywheel, an executive agenda for CEOs. The report is worth reading in full.

It's also worth rereading. Because every sentence, applied to enterprise B2B rather than consumer B2C, multiplies in complexity, in value, and in the time required to build it.

McKinsey's projections cover B2C: €600 billion in European e-commerce by 2029, 5-7% annual growth, a competitive shift from front-end attributes to back-end reliability. These are serious numbers.

Italian digital B2B, according to Politecnico di Milano's Observatory, is already four times the size of B2C. European B2B applies the same ratio. What McKinsey describes as "shifts" — agentic commerce, the fusion of media and commerce, retail media as a margin engine, omnichannel reinvented through intelligence, competitive pressure from Chinese disruptors — is already in motion across B2B. And with substantially higher economic consequences.

In this article we take McKinsey's five shifts and retranslate them for the enterprise B2B landscape. When the two worlds converge — and they're already converging — the difference between those who built the right Ecommerce Experience Platform and those who patched together a series of tactical solutions will show up in two or three years of competitive advantage.

Shift 1 — Agentic commerce is already mainstream in B2B. It just doesn't go by that name.

McKinsey estimates that by 2030 between $3 and $5 trillion of global B2C revenue will flow through agentic commerce — AI systems that research, evaluate, and purchase on behalf of consumers. 38% of European consumers already use gen AI tools to research products.

In B2B, agentic commerce has been running in production for years. It just doesn't go by that name. It goes by "automated reordering." By "ERP systems calling supplier APIs." By "procurement rules that generate orders when stock drops below threshold." By "sales reps using CRM to recalibrate discounts and terms in real time."

A B2B buyer in a procurement role can already delegate to AI: generating an RFP, comparing incoming offers, validating supplier contracts, planning seasonal reorders. Each of these tasks removes human hours — and transfers decision power to the algorithm that orchestrates them.

The difference from B2C is that a B2B AI agent operates within much stricter operational constraints: master contracts with volume clauses, payment terms that vary per customer, price lists with time-limited validity, batch availability with concurrent reservations, documentary compliance requirements (electronic invoicing, delivery notes, product certifications).

A B2B agentic AI that ignores even one of these constraints isn't useful. It's a problem. The infrastructure underneath needs to expose every one of those constraints in a machine-readable form, updated in real time, and consistent across channels. In practical terms: normalized shared catalog, contextual pricing engine, ERP integrated as system of record, availability check with reservation lock.

It's no coincidence that the average 6:1 ratio between buyers and customer companies on the Rewix platform maps exactly to what McKinsey calls "multi-role delegation." A customer company doesn't have one buyer. It has six operational roles interfacing with the platform — each one potentially supported by its own AI agent.

Anyone who wants to compete for being selected by a buyer's agent in 2027 has to ensure their data is readable, reliable, and consistent in 2026. There's no way to rebuild that foundation in a hurry.

How to prepare your B2B catalog for agentic commerce channels →

Shift 2 — Commerce and media have been converging in B2B for twenty years. "Content-driven B2B" is already verified in the numbers.

McKinsey describes the shift as "commerce becomes media, media becomes commerce": TikTok Shop as storefront, content as showcase, feed-based shopping.

In enterprise B2B, this convergence started much earlier and quietly. LinkedIn is the #1 channel for qualified B2B lead generation in Europe. Serious technical blogs have replaced trade shows as primary discovery. Industry newsletters get read before Gartner reports. Vertical communities — Slack, Discord, LinkedIn groups — are where B2B buyers ask peer-to-peer questions before they even open an RFP.

What McKinsey calls "media becomes commerce" in B2C is, in B2B: editorial content that generates qualified leads, nurtured through content, converting along a long consideration funnel, closing with a demo. The gap between a LinkedIn post and a B2B enterprise contract can be 18 months. But the whole span passes through content.

AI amplifies this in two directions. First: agent-to-agent negotiation, where brand and buyer AI agents negotiate through structured content (product specifications, contract terms, certifications). Second: content-as-catalog-entry, where the product spec on the site is the product spec the AI agent reads to place the order.

Catalog, content marketing, and commerce experience aren't three separate things. They're a single infrastructure of structured product representation, serving humans (who read the sheet), search engines (which index it), and AI agents (which integrate it into procurement processes) simultaneously.

In B2B, content is commerce. It's not becoming that — it's been that for a decade. What AI adds is executional scale.

Discover how the Rewix Ecommerce Experience Platform manages catalog, content and commerce as one infrastructure →

Shift 3 — Retail media already exists in B2B. It's called "platform network effect."

McKinsey describes retail media as a new structural margin lever — margins up to 10× higher than core retail, monetized attention, first-party data generating measurable ROI.

In traditional B2B, the equivalent is: brand co-op with distributors, trade marketing incentives, listing fees in printed catalogs. Traditional models with historic inefficiencies and no measurable feedback loop.

In network B2B — the kind built on platforms like Rewix — a structurally more powerful version already exists: cross-brand discovery native to the network itself. In the Rewix dataset, 22% of orders contain at least one cross-brand reference — a product found by the buyer because someone else on the same network had ordered it, not because it was actively searched.

This is B2B retail media in its purest form. The network monetizes buyer attention not with banners or sponsored placement, but with the very structure of the shared catalog — normalized taxonomies, cross-references between brands, reorder patterns that fuel discovery.

The economic model is different from the B2C retail media McKinsey describes. B2B doesn't earn from selling ad space. It earns from the network's ability to make each brand more visible to a buyer who has already shown interest in adjacent categories. It's a natural slot, not a paid one. But the economics are equally — or more — favorable, because the marginal cost of generating visibility is zero: visibility is a byproduct of how the network operates.

By 2027, most B2B platforms will treat their network effect as a P&L lever, not as an organic consequence. Those who aren't measuring it today will find out too late.

How the Rewix Ecommerce Experience Platform turns network effect into a commercial lever →

Shift 4 — B2B omnichannel isn't "reinvented by intelligence." It's invented by intelligence.

In the McKinsey article, B2C omnichannel has existed for years: consumers flowing between online, store, mobile app, customer service. AI now makes this flow dynamic in real time — pricing, inventory, promotions optimized across the entire journey.

In B2B, omnichannel hasn't been "reinvented." It's been barely invented.

Most enterprise B2B relationships still run through: dedicated web portal, sales rep on the phone, order confirmation email, follow-up calls from admin, EDI systems for recurring orders, monthly reporting dashboards. Six different channels that in 95% of cases don't talk to each other.

AI applied to B2B omnichannel doesn't optimize — it rebuilds. The buyer placing an order at 9:00 AM on the web portal needs to find the same availability the sales rep promised at 8:30 on the phone. And the finance controller checking the residual credit at 9:15 from the dashboard needs to see the order already computed with the correct pricing for that specific customer.

This is why an ERP-first architecture is Rewix's reason for existence: the ERP is the system of record, the commerce platform is the engagement layer. Operational decisions are made where the data lives, not where the buyer clicks. Every touchpoint — portal, mobile app, phone call, email — reflects the same operational reality, updated in real time.

How Rewix delivers real-time operational consistency across every touchpoint →

Shift 5 — Competitive pressure from Chinese disruptors arrived in B2B twenty years ago. What's changing is the sophistication of curation.

McKinsey describes the pressure from Chinese disruptors in B2C: Shein, Temu, TikTok Shop. Compressed delivery times, elimination of intermediaries, manufacturer-led brands moving up the value chain.

In B2B, this competition has been around for twenty years. Alibaba.com was founded in 1999. Global Sources is even older. Chinese manufacturers have been selling directly to European buyers for more than two decades. European B2B buyers have built risk management playbooks (guarantees, pre-shipment inspection, certified alternative sourcing) that B2C is only now discovering.

What's changing in B2B, driven by AI, is the sophistication of curation. European B2B buyers can access Chinese assortments, but increasingly through intermediation platforms that add quality control, compliance verifications (CE marking, REACH, GDPR on data), documentary management for European invoicing, and post-sale guarantees.

B2B platforms that integrate international supply chains with European operational standards become curated network marketplaces: they offer European buyers access to global suppliers with the commercial, contractual, and operational guarantee of the European platform. This isn't a defense against Chinese competition — it's the orchestration of it.

As we wrote in our earlier piece on the B2B e-commerce iceberg, the competitive advantage of a B2B enterprise platform lies in what sits beneath the surface — not in the price of the product on top of it.

How Rewix integrates global supply chains with European operational standards →

The B2B flywheel: same mechanisms, multiplied by the enterprise cycle

McKinsey describes a flywheel: four levers — Growth (+10-15%), Productivity (+30-50%), Margin (+3-5 points), Value Chain (+10-20%) — reinforcing each other when integrated. In B2B, each lever has a broader version.

Growth. In B2C, personalization means recommendation engine. In B2B, personalization means: a buyer accessing the portal sees their contracted price list, their reorder history, their pending orders, their residual credit, their payment terms — all in real time. It's not "next-best-offer." It's "context-perfect state."

Productivity. McKinsey cites AI copilots reducing customer care handling times by 40-60%. In B2B, where each order carries an average of 10-15 messages between buyer and sales before confirmation, the reduction of human touchpoints is proportionally larger — and the recovered value is higher because a B2B sales hour costs more than a B2C customer service hour.

Margin. In B2C, dynamic pricing optimizes at the margin for 2-5 points. In B2B, dynamic contract management — volume discounts, customer terms, raw material surcharges, validity windows — isn't optimization. It's automated execution of business rules that today are still applied manually, with an error rate measurable in margin percentage points.

Value Chain. B2B supply chain is an order of magnitude more complex than B2C: multi-warehouse, multi-batch, cross-brand, with variable lead times per product and per customer. AI applied here doesn't reduce inventory costs by 10-20%. It reduces by 40-60% the time spent managing allocation conflicts between concurrent buyers on the same batch.

The B2B flywheel compounds faster because each improvement cycle involves multiple entities (buyer, sales rep, ERP, warehouse, finance) and each feedback loop informs every other entity.

What "rewired" means in B2B

McKinsey closes their article by citing their firm's Rewired framework: build AI into the way daily decisions are made, not alongside them. Rewire workflows, don't add layers.

In enterprise B2B, "rewired" has a literal meaning. The decision-makers in a customer company operating with Rewix don't make decisions in isolation — they make them within workflows involving five or six roles, three or four systems (ERP, CRM, PIM, WMS), and multiple validation cycles.

Rewiring means reconnecting those nodes so information flows at the speed AI can process it, not at the speed at which humans pass emails between themselves.

Rewix was designed for exactly this purpose: composable architecture, integrated via API with the enterprise ecosystem, with AI as native layer and not as add-on feature. McKinsey's flywheel works when the data generated by each interaction informs the next. In B2B, that data is richer, more structured, and has a higher commercial value per interaction. The economics of rewiring in B2B are stronger precisely because each B2B "click" is worth substantially more than each B2C click.

Why native AI is the core of the Rewix Ecommerce Experience Platform →

The agenda for B2B leaders

McKinsey closes with five actions for retail CEOs. We take the same structure, applied to enterprise B2B.

1. Make your infrastructure agent-ready. Not in the sense of "clickable by an AI agent" — in the sense of machine-representable across every dimension: normalized catalog, contextual pricing, real availability, customer terms, approval workflows. Every B2B buyer who delegates a recurring reorder to an AI agent over the next 24 months will find, on the other side, the supplier platform that's ready or the one that isn't. The difference will show up in contract retention.

2. Build a content engine as part of the commerce infrastructure. In B2B, this means: the product specification isn't a marketing asset, it's a commercial asset. Every description, every attribute, every certification, every image is simultaneously input for the buyer's AI agent and content validating the relationship. The commercial catalog is the content marketing.

3. Treat the network effect as a P&L line, not a secondary benefit. The value of cross-brand discovery — in Rewix's dataset, 22% of orders contain at least one cross-brand reference — is quantifiable. It should be measured, optimized, and made visible as a strategic lever, not as an organic consequence of the network.

4. Embed AI in daily trading decisions. Not in annual planning, not in periodic review cycles. In the moment a buyer requests a quote for 500 units of a specific product on a specific delivery date, AI should have already computed the optimal combination of volume discount, batch allocation, and payment terms — which the sales rep can validate or modify, but shouldn't build from scratch.

5. Rewire the operating model. Every role that today coordinates information between different systems (sales rep asking product manager, product manager asking finance, finance asking warehouse) is a candidate for rewiring. AI doesn't eliminate these roles — it frees them for the work AI can't do: building relationships, managing exceptions, negotiating complex cases.

Assess whether your platform is ready for these five shifts →

The conclusion: infrastructure is strategy

The McKinsey article closes with this sentence: "Treat AI not as a set of tools, but as the foundation of the commercial system itself."

In B2B, this isn't an aspiration. It's a requirement.

B2B enterprise platforms that have spent eight or ten years working on their infrastructure now have the option to activate AI as a native layer. Platforms that spent the same time on UX design and front-end optimization now have to rebuild from the foundations.

Rewix belongs to the first category. Eight years of B2B commercial infrastructure. $580 million in processed transactions. 209,000 customer companies. The 6:1 ratio between buyers and companies — the mathematics of operational network effect. An Ecommerce Experience Platform with an ERP-first architecture that is AI-native because it was designed to be, not retrofitted to look like it.

And those who already built the foundations can rewire in months what those who build from scratch will take years to construct.