By 2030, between $3 trillion and $5 trillion in retail sales will be mediated, orchestrated, or influenced by AI agents, according to McKinsey. However, only 2% of organizations are truly prepared to meet the challenges of agentic commerce.
Only 2% of e-commerce companies worldwide have fully embraced industrial commerce agents. This means they have an AI agent capable of executing a financial transaction from start to finish without human intervention.
This capability opens up a world of possibilities, both for consumers and e-commerce businesses. Indeed, a growing number of shoppers are turning to generative AI tools, such as ChatGPT, Perplexity, Gemini, or Claude, to help them find and compare goods and services. Generative AI thus emerges as a genuine aid to analysis and decision-making.
However, the following steps are often still within the human sphere, namely checking the availability of the product, the reliability of the merchant site, and of course actually placing the order.
In the near future, intelligent agents will be the ones executing purchasing acts on a massive scale.
Therefore, the 2% rate of companies ready for agentic commerce seems very low considering the upcoming business opportunity. And not just in ten years. Behaviors are changing… and fast. Indeed, according to McKinsey & Company , $1 trillion in retail sales will be mediated, orchestrated, or influenced by AI agents by 2030; and that's in the US market alone. Internationally, the consulting firm's estimates range between $3 trillion and $5 trillion!
Agentique commerce: what are we talking about?
Agentic commerce refers to a new e-commerce model: the act of purchasing is no longer the sole domain of a human browsing a traditional website. It can be delegated to an autonomous AI agent! In short, we're moving from the traditional "Search & Click" to the revolutionary "Prompt & Delegate." It's a true paradigm shift!
Unlike traditional chatbots or recommendation engines, the intelligent agent is capable of:
- to interpret a complex purchase intent. For example: "I have my B2B conference in Lyon next Tuesday. Organize my travel, find me a dressy outfit suitable for the local weather, and arrange for our 200 sales brochures to be restocked directly at the hotel for Monday evening."
- to consult the catalogue, individualized prices and availability in a dynamic way.
- to negotiate or validate the terms of sale.
- to execute checkout and payment securely from end to end.
The 4 conditions of commercial agency
Transitioning to agentic commerce is not simply a matter of adding a layer of generative AI to an existing website. It requires a more profound redesign.
- Restructuring websites into autonomous components (API-first architecture): AI agents don't navigate HTML pages with a mouse; they consume APIs. Without a fully headless and decoupled architecture, machines can neither read the offer nor execute actions.
- Natively handle complex intents: Your systems must be able to understand structured natural language and infer multi-criteria contextual queries in real time, far beyond simple keyword searches.
- Unifying real-time data (PIM + ERP/OMS + CRM/CDP): Without synergy between product data (PIM), stock availability and supply chain (ERP/OMS), as well as customer history and privileges (CRM/CDP), the agent cannot finalize the order.
- Integrating automated checkpoints (Guardrails): Delegating checkout to agents implies securing financial flows: spending limits, identity verification of purchasing agents, fraud detection and programmable authorization protocols.
Agentique commerce: 3 concrete examples
As mentioned in the introduction, some players in the e-commerce ecosystem have already laid the necessary technical foundations to be “consumed” by machines and thus meet the requirements of agentic commerce.
1. Shopify and its smart selection of ready-to-buy products
In 2025, e-commerce giant Shopify launched its agentic commerce infrastructure , enabling the entire purchase journey to be completed directly via AI-powered conversational agents.
In short, the buyer simply tells their AI agent what they're looking for, such as a new pair of shoes for a hike or a birthday present for a comic book-loving teenager. On Shopify's side, the AI agents search for products and ask questions to refine the request, specifying price, size, delivery method, and other details. Shopify's agents then present a selection of products ready for purchase.
2. Nike inserts itself into conversations
Nike launched an experiment in agentic commerce in the spring of 2026, ahead of the FIFA World Cup. The sportswear company made every effort to ensure its products could be easily recommended and found through the Gemini and Google AI platforms. Most importantly, the brand's goal was for consumers to purchase its products directly during conversations with AI agents.
3. Wayfair and its machine-readable products
Following Nike's lead, furniture retailer Wayfair has been shifting towards agentic commerce since 2025, ensuring its catalog of some 30 million products from over 20,000 suppliers is readable by AI agents. In an interview published on Constellationr.com, Fiona Tan, CTO of Wayfair, explains: "Our goal is to make our catalog fully accessible on major AI platforms, so customers can shop with confidence, no matter where they start their journey." She adds: "In a world where commerce is driven by AI, retailers with a large and detailed catalog, verified supply chains, and strong technological capabilities have an advantage."
Another major signal that testifies to the growing importance of agentic commerce: the entry of Visa! The specialist in digital payments launched Agentic commerce ready in March 2026, a program to ensure that these new types of transactions can be carried out securely with the explicit consent of users.
To learn more, download our eBook Back to Agentic commerce !
