- Apr 28
- 6 min read
Agentic commerce is a new ecommerce model where autonomous AI agents, could be a large language model (LLM) like ChatGPT, Gemini or specially built AI shopping assistants, act as intermediaries to research, negotiate and complete purchases on behalf of a consumer.
Unlike traditional e-Commerce, where a buyer visits a website or online store and navigates a checkout funnel, agentic commerce happens where ever the buyer prompted the agent.
How does agentic commerce work?
The first difference between traditional and agentic commerce is the zero-click experience. You might tell your AI or LLM (large language model): "Buy me a durable pair of trail running shoes under $150 and have them here by Friday." An AI agent then scans multiple online stores, compares pricing and shipping speeds, it can then use your store payment details to buy them instantly.
Instead of the human buyer, ending up on a checkout page, the AI agent talks to the store’s backend via protocols (like ACP or UCP). The checkout then happens inside the AI's chat interface.
In advanced agentic commerce cases, a personal shopping agent can negotiate with a brand’s seller agent for a bulk discount or a better shipping rate in milliseconds. It can be used both for personal shopping but also by businesses between each other and suppliers.
Learn more about what an AI agent is.
How does agentic commerce work on Wix?
Wix is the first CMS to be an official signatory of the Agentic Commerce Protocol (ACP). On Wix, agentic commerce works in two main ways:
Selling to AI agents
Wix has re-engineered its backend so that Wix stores are legible to AI agents. Here's how:
Direct AI checkout: Through partnerships with Stripe and OpenAI, Wix products can be bought directly inside ChatGPT. A user doesn't have to visit your Wix site, they see the product and a Buy button right in their chat. But you need you eCommerce website to act as a backend or inventory of your product catalog.
Protocol interoperability: Wix sites automatically format your product data (inventory, shipping and tax) so agents from Google, Perplexity and Anthropic can read the store and complete transactions without some of the mistakes humans might make.
AI visibility: Wix provides a visibility overview tool that helps merchants track how often their online brand is being recommended and purchased by different AI models. This then helps with optimizing for agentic commerce to sell more.
Managing via agents
Wix has its own internal AI agents to help online sellers run their online stores more efficiently:
Aria: You can tell Wix's AI assistant Aria to "apply a 20% discount to all summer items" or "show me my sales analytics for the last week," and it executes the task immediately.
Juno (Front desk): An AI agent that handles customer support, prioritizing messages and suggesting on-brand responses across various channels.
Kleo (Marketing): Automatically creates and executes SEO and social media marketing plans based on your store's performance data.
Is agentic commerce the future for sellers?
While new, agentic commerce already has many pros for anyone selling online. And as a type of eCommerce its growing fast, according to Morgan Stanley, agentic shoppers could represent $190-385 billion in US eCommerce spending by 2030, potentially representing 10-20% of market share.
Conversion rates increase: AI agents don't get distracted like human buyers to, which reduces the dreaded concept of cart abandonment. When an agent lands on your online store for a specific product via an API, it's there with a specific intent to buy. So if your data matches the user’s requirements, the transaction is pretty much guaranteed.
Global revenue: AI Agents never sleep not even when it comes to shopping online. Unlike human shoppers who have to sleep or work, AI agents operate 24/7 and 365 days a year. This can mean a steady, predictable stream of revenue that isn't constrained by human constraints.
Predictive inventory management and pricing: Since agentic commerce relies on structured data, sellers can use something called Seller Agents to respond in real-time. If a competitor runs out of stock, your agent can instantly detect the demand spike and adjust your prices or bundles to capture that traffic before a human store manager could even open a laptop.
Is agentic commerce the future for buyers?
Efficiency: The most obvious win for consumers is the end of window shopping. Instead of spending three hours comparing vacuum cleaner specs and reading reviews, an AI agent analyzes thousands of data points and completes the purchase in seconds.
Hyper-personalization: Traditional algorithms guess what buyers like based on their click behavior. Agents however know much more detailed information about you (e.g., your calendar, your past preferences, your budget and even your clothing sizes) and can find products that fit your specific life needs perfectly.
Less decision fatigue: By filtering out noise and paradox of choice, agents present only the top 2-3 logical options or simply opt the best choice, freeing up a your decision making for more important tasks.
Better prices: Agents don't get tired of hunting for coupons. They can monitor price drops across the entire web 24/7 and calculate the total cost of ownership (shipping + taxes + duties) more accurately than a human can.
Agentic commerce, seller beware?
What happens to business branding? Most of your branding, the beautiful product and photography, the carefully crafted About Us story and your site design, is invisible to an AI agent. It only cares about machine-readable specs, think price, weight, materials, shipping time. This turns your products into commodities where you're forced to compete on data rather than brand emotion. The power and beauty of creating a brand can be lost in an agentic commerce world.
The data wall: When a purchase happens via an agent, like inside an LLM such as ChatGPT, the customer never visits your website. This means you lose first-party data like heatmaps, browsing behavior and the ability to upsell through traditional site pop-ups. You risk becoming a ghost supplier where the customer feels more loyalty to the AI that found the product than to your brand.
The API or die technical burden: To be visible with agentic commerce a simple website isn't enough. You need to maintain one with high speed and performance, perfectly accurate product APIs and structured schema (like ACP). If your inventory count is off by even one unit or your API takes too long to respond, the AI agents will simply blackball your store from their recommendations to avoid a failed transaction for their user.
Learn more:
Agentic commerce, buyer beware?
On top of its benefits, agentic commerce is still new and developing, meaning it comes with a learning curve and challenges:
What happens to brand discovery?: When an agent shops for you or your potential customers, you and them lose the joy of the find. You're less likely to discover a new, quirky brand or a product you didn't know you needed because the agent is optimized for logic and utility, not exploration. Although this may change as agentic commerce develops and advances.
Is there any transparency?: It can be difficult to know why an agent chose one product over another. There's a risk of pay-to-play bias, where AI companies might prioritize brands that pay for higher agent visibility rather than the product that's actually best for the user.
Security and unauthorized spending: Giving an AI agent the keys to your wallet is a massive trust leap. If an agent hallucinates or is hit by a malicious third-party site, it could theoretically make unauthorized purchases or leak sensitive financial data.
Agentic commerce use cases we've got our eye on
01. Personal shopper
This is set to be the most common consumer use case. Instead of a shopper looking for individual items, they delegate a complex goal to their personal agent.
They might say, "I’m going hiking in Zion National Park next month. I need a complete gear setup for under $800 that fits my size Medium, and make sure it arrives by the 15th."
The agent doesn't then just show a list; it checks the Zion weather forecast for May, cross-references the user’s past clothing purchases for sizing, and scans multiple retailers via the Agentic Commerce Protocol (ACP).
It then builds a bundle deal from three different stores, negotiates a first-time buyer discount at one and presents the final one-click checkout to the user.
02. Autonomous predictive replenishment
This use case removes the supply ordering step entirely, moving from a human-triggered purchase to a need-triggered one.
Lets say a small coffee shop uses an agentic system to manage its inventory. Its AI agent monitors real-time sales (via the POS system) and local events (like a nearby festival that will increase foot traffic). Seeing that oat milk will run out 48 hours earlier than usual, the agent autonomously contacts three local suppliers. But doesn't just reorder it, it also negotiates based on the urgency, finds the supplier with the lowest emergency delivery fee and then pays. The owner simply gets a notification: "Oat milk restock arriving tomorrow at 6:00 AM to cover the festival rush."
03. Customer support
For example, an expensive electronic good is ordered but then delayed in shipping shipping because of a storm along the route. An agent monitoring the shipment detects the problem. It looks at the merchant’s policy and sees the customer has Gold Status with the brand. Without waiting for a customer complaint, the agent:
Reroutes a duplicate item from a closer physical store.
Messages the customer: "Your original shipment is delayed by a storm, so I’ve intercepted a new unit from a local branch. It will still arrive today by 5:00 PM."
Issues a 10% hassle-free credit to the customer’s account for the inconvenience.
