Agentic Score: How AI Interacts With You
Your Agentic score measures whether AI agents can take actions on your behalf -- browse products, make purchases, complete tasks. It is the measure of how ready your site is for the age of autonomous AI.
Score Range
What We Check
Your Agentic score evaluates six capabilities that determine whether AI agents can meaningfully interact with your site beyond simple content retrieval.
MCP Manifest (/mcp.json)
+20 pointsA Model Context Protocol server declaration at your domain root. MCP is the emerging standard for how AI agents discover and interact with external services. The manifest declares what tools and capabilities your site exposes to AI agents.
Read implementation guideMCP Endpoints
+15 pointsFunctional API endpoints that AI agents can call to perform actions on your site. These go beyond the manifest declaration -- they are the actual working interfaces that let agents browse products, retrieve data, or execute tasks.
Read implementation guideMCP SSE (Real-time)
+15 pointsServer-Sent Events endpoints for live, streaming communication between AI agents and your site. SSE enables real-time updates -- inventory changes, price updates, order status -- without requiring the agent to poll repeatedly.
Read implementation guideagents.json
+15 pointsAn agent discovery and capability declaration file that tells AI systems what autonomous agents can do on your site. It defines the agent's identity, supported actions, authentication requirements, and interaction protocols.
Read implementation guideCommerce Capabilities
+20 pointsProduct Schema markup, Payment Intent declarations, and Checkout API endpoints that enable AI agents to browse your catalog, compare products, and complete purchases on behalf of users. The building blocks of agentic commerce.
Read implementation guideA2A Protocol
+15 pointsAgent-to-Agent communication protocol support that allows AI agents from different systems to coordinate, negotiate, and transact with each other through your platform. The next frontier of multi-agent collaboration.
Read implementation guideWhy Agentic Matters
We are entering the age of AI agents. These are not chatbots that answer questions -- they are autonomous systems that can browse the web, compare products, negotiate prices, and complete purchases on behalf of users. OpenAI, Google, and Anthropic are all building agent capabilities, and the infrastructure for agent-to-site interaction is being standardized right now.
The Model Context Protocol (MCP) is emerging as the standard for how AI agents discover and interact with external services. Sites that implement MCP manifests, expose functional API endpoints, and support real-time communication via SSE are positioning themselves to capture the first wave of agentic commerce.
Consider the implications: when a user tells their AI assistant to "find me the best project management tool under $20/month and sign me up for a trial," the agent needs to browse product pages, compare pricing, and complete a signup flow. If your site exposes these capabilities through structured protocols, you are in the running. If you do not, the agent will transact with your competitor.
Agentic readiness is not just about e-commerce. Service businesses, SaaS platforms, content publishers, and even local businesses benefit from agent-accessible interfaces. AI agents will book appointments, request quotes, download resources, and integrate with workflows. The businesses that make these actions frictionless for agents will win disproportionate market share.
How to Improve Your Agentic Score
Create an MCP manifest
Place an mcp.json file at your domain root declaring your site's available tools and capabilities. This is the entry point for AI agent discovery -- without it, agents cannot find your APIs.
Expose functional MCP endpoints
Build API endpoints that implement the actions declared in your manifest. Start with read-only operations (product listings, pricing, availability) before adding write operations (cart, checkout, booking).
Add agents.json
Create an agent discovery file that declares your site's agent capabilities, supported authentication methods, and interaction protocols. This helps AI systems understand what they can do on your site.
Implement Product Schema markup
Add structured product data (pricing, availability, reviews, specifications) that AI agents can parse and compare across sites. Schema.org Product markup is the foundation of agentic commerce.
Add SSE streaming endpoints
Implement Server-Sent Events for real-time data like inventory updates, price changes, and order tracking. SSE allows agents to maintain live connections instead of polling.
Explore A2A protocol support
For advanced use cases, implement Agent-to-Agent communication that allows AI agents from different platforms to coordinate through your site. This is the cutting edge of multi-agent commerce.
Related Articles
See Your Agentic Score
Scan your domain to discover how ready your site is for AI agent interaction. Get actionable recommendations to become agentic-ready before your competitors.
Scan Your Domain