What Is Agent Readiness? The Complete Guide
Agent readiness is how prepared a business is to be discovered, understood, and transacted with by AI agents. It is the new competitive dimension that determines whether your business participates in the agent economy — or gets left behind.
This guide covers everything you need to know: why agent readiness matters, the 7-level ARL framework, the 6-step agent journey, 9 scoring dimensions, and the $6.2B market opportunity. Based on data from scanning 238+ real businesses.
In This Guide
Why Agent Readiness Matters
AI agents are learning to shop, book, and pay. Not in theory — right now. When a user tells ChatGPT “find me a plumber in Austin who can come today,” the AI agent goes looking for structured data it can parse. It checks for Schema.org markup, reads Google Business Profiles, scans for API endpoints, and looks for machine-readable service listings. The businesses with structured data get recommended. The businesses without it get skipped.
This is the same pattern that played out with search engines in the 2000s. When Google became the primary way people found businesses, the businesses with websites got found and the businesses without them disappeared. SEO became a multi-billion-dollar industry overnight. Agent readiness is the next version of this pattern — but instead of optimizing for humans who click search results, you are optimizing for AI agents who act on behalf of humans.
The difference is speed. SEO took 10 years to become table stakes. Agent readiness will take 2-3 years. AI agents are being deployed by every major tech company simultaneously — Apple, Google, Amazon, Microsoft, Anthropic, OpenAI. Each is building agents that need to interact with businesses. The businesses that are ready will capture this wave. The businesses that are not will lose customers they never even know existed.
Agents are learning to transact
AI agents can already search, compare, and in some cases book and pay. Each month brings new capabilities.
Invisible businesses lose
If an agent cannot find your business in structured data, it recommends your competitor. You never know what you lost.
The window is closing
First movers in SEO dominated for a decade. First movers in agent readiness will have the same compounding advantage.
The 6-Step Agent Journey
Every AI agent follows a predictable 6-step journey when interacting with a business. Most businesses fail at step 1 — agents cannot even find them. Understanding where your business breaks in this journey is the key to improving your agent readiness.
FIND
Can the agent discover this business exists?
The agent needs to know the business exists and what it does. This requires structured data: Schema.org markup, Google Business Profile, agent-card.json, llms.txt, or inclusion in an agent-searchable registry. If the business has no machine-readable presence, the agent will never recommend it.
Pass
Agent card, Schema.org markup, or Google Business Profile with structured data.
Fail
A business with only a static HTML website and no structured metadata. Agents searching for this type of service will never find it.
UNDERSTAND
Can the agent parse what the business offers, at what price, and availability?
Once found, the agent needs to understand the specifics: what products or services are available, what they cost, and whether they are in stock or have open time slots. This data must be structured — JSON, XML feeds, or Schema.org — not embedded in HTML paragraphs or PDF documents.
Pass
Structured product catalog, service list, or menu with prices, descriptions, and availability.
Fail
A restaurant whose menu is a scanned PDF image. The agent can see the restaurant exists but cannot determine the dishes, prices, or allergen information.
SIGN UP
Can the agent create an account or start a relationship programmatically?
Many businesses require signup, registration, or account creation before transacting. If this process is purely GUI-based (CAPTCHA, email verification, identity upload), an agent cannot complete it. Agent-ready signup means API-accessible account creation or token-based access.
Pass
API-based signup, OAuth flow, or API key provisioning without human-only verification.
Fail
A SaaS that requires clicking a verification email, completing a CAPTCHA, and uploading a photo ID. The agent cannot proceed past the first screen.
CONNECT
Can the agent authenticate and establish a working connection?
The agent needs credentials (API keys, OAuth tokens) and a stable endpoint to interact with. The API must return structured responses — JSON with consistent error formats, not HTML error pages. Rate limiting, versioning, and authentication must be predictable and documented.
Pass
REST or MCP endpoints with API key auth, JSON responses, and documented error codes.
Fail
An API that returns HTML error pages, has no rate limit headers, and requires cookie-based session auth that expires every 15 minutes.
USE
Can the agent perform the core action: book, order, query, or transact?
This is where value is created. The agent needs to be able to do something: place an order, book an appointment, run a query, or trigger a workflow. The API must support the primary use case end-to-end with programmatic input and structured output.
Pass
Functional API endpoints for the core business action with input validation and confirmation responses.
Fail
A booking system that accepts the reservation request but only confirms via email 24 hours later with no API-queryable status.
PAY
Can the agent complete payment programmatically?
The final step. If the transaction requires payment, the agent must be able to pay without a human entering credit card details into a web form. This means Stripe, Square, or similar programmatic payment processing. Some business models skip this step (freemium, invoice later), but most agent-driven transactions end in payment.
Pass
Stripe, Square, or direct payment API. Pre-authorized payments, wallet-based billing, or x402 micropayments.
Fail
A checkout flow that requires filling out a web form with credit card numbers, billing address, and a CAPTCHA. No agent can complete this.
Most businesses fail at step 1 or 2. Our data shows that 63% of businesses we scan cannot even be found by agents (step 1), and another 20% are findable but have no structured offerings for agents to read (step 2). Only 8% of businesses are ready for agents to actually transact with them at step 5 or 6.
The 7 Agent Readiness Levels (ARL)
The Agent Readiness Level (ARL) framework measures where a business falls on the spectrum from completely invisible to fully interoperable. Each level is cumulative — you cannot be ARL-4 without meeting all requirements for ARL-0 through ARL-3.
ARL-0: Dark
Score 0-19Invisible to AI agents — no structured data, no API, no machine-readable presence.
A local plumber with a brochure website. Phone number is an image. Hours are on Facebook.
ARL-1: Discoverable
Score 20-34Agents can find the business and understand what it does at a basic level.
A restaurant with a Google Business Profile — hours, cuisine, and location are structured.
ARL-2: Readable
Score 35-49Agents can read offerings, pricing, and availability in structured formats.
An e-commerce store with Schema.org product markup — prices, stock, and reviews are machine-readable.
ARL-3: Bookable
Score 50-59Agents can initiate transactions — book, order, or request a quote programmatically.
A SaaS with self-serve signup and API keys. An agent can create an account and start using it.
ARL-4: Transactable
Score 60-69Agents can complete the full transaction cycle: create, pay, track, modify, and cancel.
Stripe — agents can create payments, manage subscriptions, issue refunds, and track status.
ARL-5: Autonomous
Score 70-74Agents can manage the ongoing relationship — reorder, optimize, escalate, and renew.
AWS — agents provision resources, auto-scale, optimize costs, and manage billing across lifecycle.
ARL-6: Interoperable
Score 75+The business runs its own agent that communicates with customer agents via A2A and MCP.
A restaurant whose agent negotiates group rates with a corporate travel agent, fully automated.
The 9 Scoring Dimensions
The Agent Readiness Score is a weighted composite of 9 dimensions. Each dimension measures a different aspect of how well a business can interact with AI agents. The weights reflect real-world importance — API Quality (15%) matters more than Pricing Transparency (5%) because agents need callable endpoints before they need to compare prices.
Can agents find the business? Agent cards, Schema.org, llms.txt, AGENTS.md, registry presence.
Are there clean, callable API endpoints with consistent JSON responses and proper error codes?
Can an agent create an account, get credentials, and start using the service programmatically?
Is pricing structured and machine-readable, not buried in "Contact Sales" buttons?
Can agents pay programmatically via Stripe, Square, x402, or similar payment APIs?
Are product/service listings structured with consistent schemas, descriptions, and availability?
TLS, proper authentication, rate limiting, secure credential management, no leaked secrets.
Uptime, response time, consistent behavior. Can agents depend on this service?
SDKs, error handling, tracing, webhooks. How easy is it for agents to integrate deeply?
Score Tiers
Platinum
90+
Gold
75-89
Silver
60-74
Bronze
40-59
Not Scored
0-39
How Agent Readiness Differs from SEO
SEO and agent readiness solve related but fundamentally different problems. SEO optimizes for search engine crawlers that rank pages for humans to click. Agent readiness optimizes for AI agents that act on behalf of humans — finding, comparing, booking, and paying without the human ever visiting a website.
The key difference: with SEO, the human makes the final decision and takes the action (clicking, reading, buying). With agent readiness, the agent makes the recommendation AND takes the action. The business that is agent-ready gets booked directly. The business that only has good SEO gets listed in results that the human may or may not click on — and even if they do, they still have to complete the transaction manually.
| Dimension | SEO | Agent Readiness |
|---|---|---|
| Audience | Search engine crawlers | AI agents acting for users |
| Goal | Rank high in search results | Be selected and transacted with by agents |
| Outcome | Human clicks a link and visits site | Agent completes the transaction directly |
| Format | HTML, meta tags, backlinks | JSON APIs, MCP tools, agent cards, llms.txt |
| Key metric | Click-through rate | Transaction completion rate |
| Timeframe | Became critical 2005-2010 | Becoming critical 2025-2028 |
| Revenue model | Traffic to conversion funnel | Direct agent-to-business transactions |
They are complementary, not competing. A business should have both strong SEO and high agent readiness. SEO captures the users who search and browse. Agent readiness captures the users who delegate to AI agents. As agent usage grows, the share of business that flows through agents (rather than search results) will increase every year.
How Agent Readiness Differs from Web Accessibility
Web accessibility (WCAG) and agent readiness both involve making content available to non-traditional consumers — but they serve different audiences and require different implementations.
Accessibility ensures that humans with disabilities can perceive, navigate, and interact with websites. It focuses on screen readers, keyboard navigation, color contrast, and alt text. Agent readiness ensures that AI software agents can discover, understand, and transact with businesses. It focuses on structured data, APIs, machine-readable pricing, and programmatic payment.
An accessible website might have perfect ARIA labels and screen reader support, but still be completely invisible to AI agents because it has no structured product data, no API, and no machine-readable pricing. Conversely, a business with excellent agent readiness (clean APIs, MCP server, structured data) might have a website that fails basic accessibility audits.
Web Accessibility (WCAG)
- Audience: Humans with disabilities
- Tools: Screen readers, keyboard, switches
- Standards: WCAG 2.1, ARIA, Section 508
- Focus: Perceivable, operable, understandable UI
- Legal: ADA compliance requirements
Agent Readiness (ARL)
- Audience: AI agents acting for users
- Tools: MCP clients, API callers, A2A agents
- Standards: MCP, A2A, Schema.org, OpenAPI
- Focus: Structured data, APIs, programmatic access
- Legal: No requirements yet (emerging field)
Industry Data: 238 Businesses Scanned
We have scanned 238 businesses across 15+ verticals using the AgentHermes 9-dimension scanner. The data paints a clear picture: most businesses are not ready for the agent economy.
Score Distribution
Top Performers
Supabase
69
Silver
Vercel
69
Silver
Slack
68
Silver
Stripe
68
Silver
Even the best score only 69. No business in our database has achieved Gold (75+) or Platinum (90+) status. The highest-scoring businesses — major tech platforms with mature APIs — still lack agent-native features like A2A agent cards and MCP servers. This is an industry in its earliest days. The opportunity to lead is wide open.
The $6.2B Market Opportunity
There are 33 million small businesses in the United States alone. Globally, the number exceeds 400 million. Currently, zero of them have MCP servers. Zero have A2A agent cards. Zero are fully interoperable with AI agents.
The market for agent readiness infrastructure is conservatively estimated at $6.2B annually in the US alone. This includes the SaaS revenue from providing agent readiness tools ($99-499/month per business), the per-transaction fees from agent-mediated bookings and purchases, and the enterprise consulting revenue from helping larger organizations build agent-ready architectures.
This mirrors the SEO tools market ($80B+ industry) but with a critical difference: agent readiness has a much clearer ROI. With SEO, the connection between “higher ranking” and “more revenue” is indirect and hard to measure. With agent readiness, every agent-mediated transaction is directly attributable. You can measure exactly how many bookings, orders, and payments came through agent channels.
33M
Small businesses (US)
Zero have MCP servers. Each is a potential customer for agent readiness infrastructure.
$6.2B
Annual market (US)
SaaS subscriptions + per-transaction fees + enterprise consulting across all verticals.
50+
Verticals addressable
Restaurants, home services, healthcare, legal, auto, real estate, fitness, beauty, and more.
This is a land-grab moment. The businesses and platforms that establish agent readiness infrastructure now will own the relationships with millions of businesses as the agent economy matures. Just as Shopify captured e-commerce infrastructure and Square captured payment infrastructure, the agent readiness layer will be owned by whoever moves fastest.
How to Check Your Agent Readiness Score
Checking your Agent Readiness Score is free and takes about 10 seconds. The AgentHermes scanner probes your business across all 9 dimensions, calculates your composite score, assigns your ARL level, and provides specific recommendations for improvement.
Enter your URL
Go to agenthermes.ai/audit and enter your business website.
Run the scan
9-dimension scan probes discovery, API, auth, pricing, payment, data, security, reliability, and agent experience.
Get your score
See your composite score (0-100), ARL level (0-6), tier, and dimension breakdown.
Take action
Follow the prioritized recommendations to improve each dimension and advance your ARL.
Your score page is permanent — accessible at agenthermes.ai/score/yourdomain.com. Share it with your team, your agency, or your board. Re-scan any time to track improvement.
Frequently Asked Questions
What is agent readiness?
Agent readiness is a measure of how prepared a business is to be discovered, understood, and transacted with by AI agents. It covers 9 dimensions from API quality to payment processing, scored on a 0-100 scale. A high agent readiness score means AI agents can find your business, understand your offerings, and complete transactions without human intervention.
Why does agent readiness matter?
AI agents are learning to shop, book, and pay on behalf of users. When a user asks an AI agent to find a plumber, book a restaurant, or compare SaaS tools, the agent looks for businesses with structured, machine-readable data. Businesses that are not agent-ready will be invisible to this growing channel — losing customers to competitors that agents can actually interact with.
How is agent readiness different from SEO?
SEO optimizes for search engine crawlers that display results for humans to click. Agent readiness optimizes for AI agents that act on behalf of humans — finding, comparing, booking, and paying without the human ever visiting a website. SEO gets you ranked. Agent readiness gets you booked.
How is agent readiness different from web accessibility?
Web accessibility (WCAG) ensures humans with disabilities can use your website. Agent readiness ensures AI agents can use your business programmatically. Accessibility is about screen readers and keyboard navigation. Agent readiness is about APIs, structured data, and machine-readable pricing. Both matter, but they solve different problems.
What are the 7 ARL levels?
ARL (Agent Readiness Level) is a 7-step progression: ARL-0 Dark (invisible), ARL-1 Discoverable (findable), ARL-2 Readable (structured offerings), ARL-3 Bookable (can transact), ARL-4 Transactable (full payment cycle), ARL-5 Autonomous (relationship management), ARL-6 Interoperable (agent-to-agent communication). Each level is cumulative.
What is a good Agent Readiness Score?
The average score across 238 scanned businesses is 42/100 (Bronze tier). Silver (60+) means agents can complete basic transactions. Gold (75+) means full lifecycle management. Platinum (90+) means the business is a model of agent readiness. The top-scoring businesses today are Supabase (69), Vercel (69), Slack (68), and Stripe (68).
How do I check my Agent Readiness Score?
Go to agenthermes.ai/audit and enter your business URL. The scanner probes 9 dimensions in about 10 seconds and gives you a score, an ARL level, and specific recommendations. The scan is free and unlimited.
What is an MCP server and why does it matter?
MCP (Model Context Protocol) is the standard for AI agents to call business tools. An MCP server exposes your business capabilities — like booking, ordering, or querying — as structured tools that any AI agent can call. Without an MCP server, agents must scrape your website. With one, they interact cleanly and reliably.
How long does it take to become agent-ready?
With AgentHermes, you can go from ARL-0 to ARL-3 in 60 seconds. Enter your business details, and the platform generates your MCP endpoint, agent card, llms.txt, and registry listing automatically. Deeper integration (connecting your real-time inventory or booking system) takes longer but starts delivering value immediately.
Does agent readiness replace my website?
No. Agent readiness adds a machine-readable layer on top of your existing business. Your website continues to serve human visitors. Agent readiness makes the same information available to AI agents in a structured format they can act on. Think of it as a second front door — one for humans (your website) and one for agents (your MCP server and agent card).
How agent-ready is your business?
Get your free Agent Readiness Score in 60 seconds. See your score across all 9 dimensions, your ARL level, and exactly what to fix first.