Skip to main content
500 Businesses ScannedEntry-Level Guide

Is Your Business Invisible to AI Agents?

We scanned 500 businesses. The average Agent Readiness Score was 43 out of 100. 40% scored so low they are effectively invisible — AI agents cannot find them, cannot understand them, and cannot send them customers.

If your business is not ready for AI agents, you are already losing customers to competitors that are. Here is what our data shows, why it matters, and what you can do about it in 60 seconds.

AH
AgentHermes Research
April 15, 20269 min read

The New Way Customers Find Businesses

Something fundamental is changing in how people find and choose businesses. For twenty years, the pattern was simple: a customer searches Google, clicks a few links, reads some reviews, and makes a decision. Your job was to show up in those search results. That was SEO, and it worked.

Now there is a new pattern. Instead of searching Google, a growing number of people are asking AI agents to do the work for them. They say “find me a plumber who can come today,” or “book a table for four at a quiet Italian restaurant near me,” or “compare project management tools under $20 a month.” The AI agent goes out, finds businesses, evaluates them, and either recommends one or completes the transaction directly.

Here is the problem: AI agents do not browse websites the way humans do. They do not look at your beautiful homepage, read your about page, or scroll through your testimonials. They look for structured, machine-readable data — things like APIs, structured product catalogs, agent cards, and standardized protocol endpoints. If your business does not have these things, the agent skips you entirely. It does not even know you exist.

We wanted to know how many businesses are actually ready for this shift. So we scanned 500 of them.

What 500 Scans Told Us

We used the AgentHermes scanner to evaluate 500 businesses across 9 dimensions that matter to AI agents: Discoverability, API Quality, Onboarding, Pricing Transparency, Payment, Data Quality, Security, Reliability, and Agent Experience. Each business received a score from 0 to 100. The results were sobering.

500
businesses scanned
43/100
average score
40%
completely invisible
0
scored Platinum

Score Distribution Across 500 Businesses

Unaudited (below 40)198 businesses (40%)

Invisible to agents

Bronze (40-59)250 businesses (50%)

Visible but barely functional

Silver (60-74)~50 businesses (10%)

Agents can transact

Gold (75-89)1 business (0.2%)

Resend (score: 75)

Platinum (90+)0 businesses (0%)

Nobody

The bottom line: 90% of businesses scored Bronze or below. That means 9 out of 10 businesses are either completely invisible to AI agents or so poorly structured that agents cannot do anything useful with them. Only one business out of 500 — Resend, a developer-focused email platform — reached Gold status with a score of 75. Zero businesses reached Platinum. The opportunity to lead is wide open.

The 6-Step Agent Journey

To understand why so many businesses fail, you need to understand how AI agents interact with businesses. Every agent follows a predictable 6-step journey. Think of it like a customer walking into a store — except the customer is software, and it moves at the speed of an API call.

If your business breaks at any step, the agent stops and moves on to a competitor that does not break. There are no second chances — agents do not “come back later.”

Step 1

FIND

Can the agent discover that your business exists?

The agent searches for businesses matching a user request. It looks for structured data: Schema.org markup, Google Business Profiles, agent cards, llms.txt files, or entries in agent registries. If your business has none of these, the agent literally cannot see you. You are a ghost.

Step 2

UNDERSTAND

Can the agent figure out what you offer and what it costs?

Once found, the agent needs to parse your services, products, pricing, and availability. This data must be structured — JSON, XML, or Schema.org markup — not buried in paragraphs of text, scanned PDF menus, or "call for a quote" pages. If the agent cannot read your offerings in a structured format, it cannot compare you to competitors.

Step 3

SIGN UP

Can the agent create an account or start a relationship?

Many businesses require registration before transacting. If signup requires CAPTCHA, email verification, phone calls, or uploading documents, an agent cannot complete it. Agent-ready signup means API-accessible account creation or token-based access that software can handle.

Step 4

CONNECT

Can the agent establish a working technical connection?

The agent needs API endpoints that return structured responses. Not HTML pages. Not PDF downloads. Clean JSON with consistent error codes, proper authentication, and predictable behavior. MCP servers, OpenAPI specs, and well-documented REST APIs all work here.

Step 5

USE

Can the agent perform the core action — book, order, or query?

This is where value is created. The agent needs to complete the primary task: place an order, book an appointment, run a search, or request a quote. The action must work end-to-end with programmatic input and a structured confirmation response. A booking system that only confirms via email 24 hours later is not agent-ready.

Step 6

PAY

Can the agent complete payment without a human typing card numbers?

The final step. If the transaction requires payment, the agent must pay programmatically — through Stripe, Square, or a similar payment API. A checkout page that requires a human to fill in credit card fields, a billing address, and a CAPTCHA is a dead end for any agent.

Where Most Businesses Fail

Here is the uncomfortable truth from our 500-business scan: most businesses fail at Step 1. They never even get found. The agent searches for a plumber, a restaurant, or a SaaS tool, and your business simply does not appear in the structured data the agent is reading.

This is not because your business is bad. It is because your business was built for humans, not for software. Your website is designed for people who can read paragraphs, look at photos, and navigate menus. AI agents do not do any of that. They look for machine-readable signals — and most businesses have none.

What AgentHermes Detects (and Most Businesses Lack)

MCP servers

The protocol for AI agents to call your business tools

Nearly zero businesses

A2A protocol

Agent-to-agent communication standard

Nearly zero businesses

agent-card.json

Machine-readable identity file for your business

Nearly zero businesses

llms.txt

Instructions for AI models about your business

Extremely rare

AGENTS.md

Agent-specific documentation file

Extremely rare

OpenAPI specs

Standardized API documentation

Only tech companies

Schema.org markup

Structured data on your website

Some businesses

Platform APIs

Shopify, WooCommerce, or Square integrations

E-commerce only

The pattern is clear. The signals that matter most to AI agents — MCP servers, agent cards, A2A protocol support — are the ones that almost nobody has. Even basic structured data like Schema.org markup is missing from a large percentage of business websites. This is not a technology gap. It is an awareness gap. Most businesses do not even know these signals exist.

The ARL Framework: 7 Levels of Agent Readiness

To help businesses understand where they stand and what to work toward, we created the Agent Readiness Level (ARL) framework. Think of it like a fitness score for the agent economy — it tells you exactly how prepared your business is, from completely dark to fully autonomous.

Each level is cumulative. You cannot jump to ARL-4 without meeting all requirements for levels 0 through 3. The good news: moving from ARL-0 to ARL-2 is straightforward and can happen in minutes, not months.

ARL-0: Dark

Score 0-19

Completely invisible to AI agents. No structured data, no API, no machine-readable presence whatsoever. This is where 40% of the businesses we scanned sit today.

ARL-1: Visible

Score 20-34

Agents can find the business and understand what it does at a basic level. Has some structured data like a Google Business Profile or Schema.org markup, but cannot transact.

ARL-2: Described

Score 35-49

Agents can read structured offerings, pricing, and availability. The business has machine-readable catalogs or menus, but no way for agents to take action.

ARL-3: Connected

Score 50-59

This is the revenue inflection point. Agents can initiate transactions — book, order, or request a quote programmatically. Has an API or MCP server that agents can call.

ARL-4: Automated

Score 60-69

Agents can complete the full transaction cycle: create, pay, track, modify, and cancel. End-to-end programmatic commerce with no human intervention required.

ARL-5: Autonomous

Score 70-89

Agents manage the ongoing relationship — reorder, optimize, escalate, and renew. The business operates in the background while agents handle customer interactions.

ARL-6: Interoperable

Score 90+

The business runs its own agent that communicates with customer agents via A2A and MCP. Full agent-to-agent commerce with no human in the loop.

The $6.2B Gap Nobody Is Filling

There are 33 million small businesses in the United States. As of today, effectively zero of them have MCP servers. Zero have A2A agent cards. Zero are fully interoperable with AI agents. This is not a rounding error — it is a gaping hole in the infrastructure of the agent economy.

Meanwhile, every major technology company is building AI agents that need to interact with businesses. Apple, Google, Amazon, Microsoft, Anthropic, OpenAI — all of them are deploying agents that search for, compare, and transact with businesses on behalf of users. These agents are ready to send customers. But they have nowhere to send them.

This gap represents a conservatively estimated $6.2 billion annual opportunity. The businesses that close this gap first do not just get more customers from agents — they become the only businesses agents can find. In a world where AI agents are making purchasing decisions, being the only visible option is not just an advantage. It is a monopoly on that customer request.

33M

US small businesses

Zero have MCP servers. Each one is a potential participant in the agent economy.

$6.2B

annual market gap

SaaS subscriptions, per-transaction fees, and enterprise consulting across all verticals.

0

Platinum-ready businesses

Out of 500 scanned, not a single business achieved full agent interoperability.

This is the SEO moment of the 2020s. When Google became the dominant way people found businesses in the early 2000s, the businesses with websites won and the businesses without them disappeared. Agent readiness is the next version of this shift — but it is moving faster. The businesses that become agent-ready now will have a compounding advantage that late movers will struggle to overcome. Explore what agent readiness looks like in your industry.

How to Check Your Score in 60 Seconds

The first step is knowing where you stand. The AgentHermes scanner evaluates your business across all 9 dimensions, calculates your composite score, assigns your ARL level, and tells you exactly what to fix first. It takes about 60 seconds and it is completely free.

1

Enter your URL

Go to agenthermes.ai/audit and enter your business website or domain.

2

Scan runs

The scanner probes all 9 dimensions: discovery, APIs, onboarding, pricing, payment, data, security, reliability, agent experience.

3

Get your score

See your composite score (0-100), ARL level (0-6), tier (Bronze through Platinum), and per-dimension breakdown.

4

Fix what matters

Follow prioritized recommendations to improve each dimension. Start with discovery — it is the highest-impact fix for most businesses.

Your score page is permanent and shareable — accessible at agenthermes.ai/score/yourdomain.com. Send it to your marketing team, your web developer, or your agency. Re-scan any time to track your improvement as you make changes.

Remember: the average business scores 43 out of 100. Even small improvements can move you ahead of 90% of your competition. And if you are in a local market, the first business in your category to become agent-ready will capture all the agent-driven traffic — because agents recommend the businesses they can interact with, and right now, that is almost nobody.

Frequently Asked Questions

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 60 seconds and gives you a score out of 100, an ARL level, and specific recommendations for improvement. The scan is completely free.

What does it mean if my business is "invisible" to AI agents?

It means AI agents cannot find your business when searching on behalf of users. If someone asks ChatGPT, Claude, or any AI assistant to find a business like yours, your business will not appear in the results. Agents rely on structured data, APIs, and machine-readable information. If your business only has a basic website with no structured data, you are invisible to this growing channel.

Do AI agents actually send customers to businesses today?

Yes, and it is growing rapidly. Millions of people already use AI assistants like ChatGPT, Claude, Google Gemini, and Siri for daily tasks including finding and comparing businesses. As these agents gain the ability to book, order, and pay, the businesses they can interact with will capture this traffic. The businesses they cannot find will lose it.

How is agent readiness different from having a good website or SEO?

SEO helps humans find your website through Google search results. Agent readiness helps AI agents find, understand, and transact with your business programmatically. A beautifully designed website with great SEO can still score zero on agent readiness if it has no structured data, no API, and no machine-readable pricing. You need both, but agent readiness is the new competitive frontier.

What is the fastest way to improve my agent readiness score?

Start with discoverability. Add structured data (Schema.org markup) to your website, claim and complete your Google Business Profile, and create an llms.txt file. These steps alone can move you from ARL-0 (Dark) to ARL-1 (Visible). For a bigger jump, connect your business through AgentHermes to get an MCP server, agent card, and registry listing automatically.


Stop being invisible

500 businesses scanned. Average score: 43/100. 90% are Bronze or below. Find out where your business stands — and what to fix first.


Share this article: