E-Commerce Agent Readiness: Why Shopify and Square Score Under 30
We scanned hundreds of e-commerce businesses across Shopify, WooCommerce, and Square. The average Agent Readiness Score is 28 out of 100 — the lowest of any major tech-enabled vertical. Square scored just 8. Here is why online stores built for human shoppers are failing AI agents.
The E-Commerce Scorecard
Out of 500 businesses scanned by AgentHermes, e-commerce is the worst-performing tech-enabled vertical. These are businesses that already have websites, payment processing, and product databases — yet they score lower than restaurants and dentists on agent readiness.
The paradox: E-commerce is the most digitized business category — online stores, payment APIs, product databases, inventory systems. Yet it scores 28/100 on agent readiness because all of that infrastructure was built for humans clicking buttons, not agents making API calls.
Square Scored 8: The Lowest Major Platform
Square is a $40 billion company that processes payments for millions of businesses. It has APIs, SDKs, and a developer platform. And it scored 8 out of 100 on agent readiness — the lowest score of any major technology company in our database. Here is the dimension-by-dimension breakdown.
D1 Discoverability
2/12No public API endpoints for agents to discover. No agent-card.json, no llms.txt, no structured product feeds accessible without authentication.
D2 API Quality
0/15Square APIs exist but are entirely behind OAuth with no public endpoints. An agent cannot even see what products a Square store sells without pre-authorized credentials.
D3 Onboarding
1/8Phone-only support for developer accounts. No self-service API key generation for agents. Manual approval process that assumes a human developer.
D4 Pricing
0/5No structured, machine-readable pricing for API access. No programmatic way for an agent to understand Square pricing tiers or transaction fees.
D5 Payment
1/8Ironically, despite being a payment company, Square has no agent-accessible payment mechanism for API usage. Agents cannot self-provision or pay for access.
The core problem with Square is that its entire platform assumes a human developer will manually set up OAuth credentials, read documentation, and build a custom integration. There is no path for an AI agent to autonomously discover a Square-powered store, understand its products, and complete a purchase. From an agent's perspective, a Square store might as well not exist.
This matters because Square powers an estimated 4 million businesses in the US alone. Every one of them is invisible to AI agents — not because they lack technology, but because that technology was never designed with autonomous discovery in mind.
Shopify: Better but Still Bronze
Shopify stores average around 30/100— better than Square's 8, but still firmly in Bronze territory. Shopify has one significant advantage: every store automatically gets public JSON endpoints.
Visit any Shopify store and append /products.json to the URL. You get a structured feed of every product with titles, descriptions, prices, images, and variants. Add /collections.jsonand you see the store's category structure. This is a massive head start on discoverability.
But discoverability alone does not make a store agent-ready. Shopify stores still fail on:
No MCP server or agent card
Agents cannot discover Shopify stores through standard agent protocols. There is no agent-card.json, no llms.txt, and no MCP endpoint.
No programmatic checkout
An agent can see products but cannot buy them. Shopify checkout requires a browser session with cookies, CAPTCHA, and human interaction.
No real-time inventory API
The public JSON endpoint does not include real-time stock levels. An agent might recommend an out-of-stock product because it cannot check availability.
No structured policies
Return policies, shipping rates, and warranty information are buried in HTML pages. Agents cannot parse these to answer customer questions.
The result: Shopify stores are partially visible but not usable by agents. An AI assistant can tell a user what products a store sells, but it cannot check if items are in stock, calculate shipping, or complete a purchase. That is the difference between a score of 30 and a score of 75.
WooCommerce: Deep APIs, Poor Discoverability
WooCommerce is the opposite of Shopify in agent readiness. It has a powerful Store API that gives programmatic access to products, orders, customers, shipping zones, tax rates, and more. An agent with API credentials can do nearly anything a human store manager can.
But WooCommerce stores average only ~25/100 because of a critical flaw: discoverability is nearly zero. Unlike Shopify, WooCommerce does not expose public JSON endpoints by default. The Store API must be explicitly enabled, and consumer keys must be generated manually. An agent discovering a WooCommerce store for the first time sees a standard WordPress site with no machine-readable product data.
WooCommerce also suffers from inconsistency. Because it is open-source and runs on WordPress, every store is configured differently. Plugins, custom themes, and varied hosting environments mean an agent cannot rely on any standard endpoint or data format across WooCommerce stores. What works on one store may return a 404 on the next.
The WooCommerce paradox: It has the deepest API capabilities of any e-commerce platform but the worst discoverability. A WooCommerce store with a properly configured Store API, an agent card, and an MCP server could theoretically outscore every Shopify store. But almost none have taken those steps.
What Agent-Ready E-Commerce Actually Looks Like
No e-commerce store in our database scores Gold (75+). But based on the scoring framework and what the top SaaS companies do right, here is the blueprint for a Gold-tier online store.
Structured product data via API
Every product has a machine-readable record with title, price, variants, stock status, images, and category. Accessible via authenticated API or public endpoint.
Real-time inventory checks
An agent can call check_inventory(product_id) and get a boolean in-stock response with quantity. No scraping product pages for "Add to Cart" button state.
Agent-accessible checkout
The store exposes create_order and process_payment tools. An agent can add items to a cart, apply discount codes, calculate shipping, and complete a purchase without a browser.
MCP server with standard tools
search_products, get_product, check_inventory, create_order, track_order, get_policies — all exposed as MCP tools that any agent can discover and call.
Agent card and discovery files
agent-card.json at /.well-known/agent-card.json, llms.txt at the root, and a registry listing so agents know this store exists and what it sells.
Structured policies
Return policy, shipping rates, warranty terms, and payment methods exposed as machine-readable data — not buried in HTML pages that agents cannot parse.
A store implementing all six components would score approximately 75-85/100 — Gold tier. That is a 3x improvement over the current e-commerce average. And the first stores to reach this level will capture agent-driven traffic that their competitors cannot even see.
AgentHermes already has adapters for all three major platforms. The /connect wizard detects whether you run Shopify, WooCommerce, or Square and auto-generates the MCP tools, agent card, and registry listing specific to your platform. What took the top SaaS companies months of engineering, you can have in minutes.
Why This Matters: The Agent Commerce Shift
E-commerce is heading toward a future where a significant percentage of purchases will be initiated or completed by AI agents. When a user tells their assistant to “reorder my vitamins” or “find me a blue wool sweater under $80 in my size,” the agent will query the stores it can access programmatically. Stores without agent infrastructure will not be in the results.
This is not hypothetical. The $6.2B MCP infrastructure gap is already being filled by platforms like AgentHermes. The stores that connect first will have a structural advantage — not just in visibility, but in conversion. An agent that can check inventory and complete checkout will close the sale. An agent that can only show a product page will lose to one that can transact.
The average e-commerce score of 28/100 is not just a data point — it is a competitive opportunity. When most stores score under 30, the first one to score 75 will dominate agent-driven commerce in its category.
Frequently Asked Questions
Why does e-commerce score so low on agent readiness?
E-commerce platforms were built for human shoppers browsing websites, not AI agents making programmatic requests. Most stores lack structured product APIs, machine-readable inventory data, and agent-accessible checkout flows. The average e-commerce Agent Readiness Score is just 28/100 because most stores fail on 7 out of 9 scoring dimensions.
How did Square score only 8 out of 100?
Square scored 8 because it has virtually no public-facing agent infrastructure. Its APIs require OAuth authentication with no public endpoints, it offers no agent-card.json or llms.txt, has phone-only developer support, no machine-readable pricing, and no way for agents to self-provision access. Despite being a major tech company, Square is nearly invisible to AI agents.
Is Shopify more agent-ready than WooCommerce?
Shopify has a slight edge in discoverability because every Shopify store automatically gets public JSON endpoints at /products.json and /collections.json. This gives agents a way to discover product data without authentication. WooCommerce has deeper API capabilities through its Store API, but stores must explicitly enable and configure it. Neither platform scores above 35 without additional agent infrastructure like MCP servers and agent cards.
What would a fully agent-ready e-commerce store look like?
A fully agent-ready store would have: structured product data with real-time inventory accessible via API, an MCP server with tools like search_products, check_inventory, and create_order, an agent-card.json for discovery, machine-readable pricing and shipping policies, automated checkout that agents can complete programmatically, and structured return/refund policies. This would score 75+ (Gold tier).
Can AgentHermes make my e-commerce store agent-ready?
Yes. AgentHermes has adapters for Shopify, WooCommerce, and Square that auto-generate MCP tools for your store. The /connect wizard detects your platform and creates agent-accessible endpoints for product search, inventory checks, and order placement. You can go from a score of 25 to 60+ in under 5 minutes.
Score your e-commerce store
Find out how your online store scores across all 9 dimensions. The scan takes 60 seconds, detects your platform automatically, and shows you exactly what to fix.