Why Most SaaS Companies Score Bronze for Agent Readiness
We scanned 500 businesses across every vertical. The average Agent Readiness Score is 43 out of 100. Half scored Bronze. Zero scored Platinum. Even Stripe — one of the best APIs ever built — only manages 68. Here is what the data says about why SaaS companies with world-class APIs still fail at agent readiness.
The Numbers: 500 Scans, One Uncomfortable Truth
We built the AgentHermes scanner to answer a simple question: if an AI agent tried to use your product today, how far would it get? Not “can a developer integrate your API” — that is a solved problem. The question is whether an autonomous agent can discover you, understand your offerings, sign up, connect, use the product, and pay for it — all without a human clicking through a UI.
After 500 scans across SaaS, e-commerce, healthcare, restaurants, and local services, the answer is clear: the agent economy is not ready. The average score is 43 out of 100, placing the typical business squarely in Bronze tier. The median is even lower. And this is not because the businesses we scanned are bad — many of them are industry leaders with excellent products and well-documented APIs.
The problem is that “great API” and “agent-ready” are not the same thing. Agent readiness is a 9-dimension problem, and most companies are only solving two or three of them.
Tier Breakdown: Where 500 Businesses Land
The distribution tells a stark story. Half of all businesses we scanned — 250 out of 500 — land in Bronze tier (40-59). These are not invisible businesses. They have websites, APIs, and sometimes excellent documentation. But they have critical gaps that prevent an AI agent from completing a full workflow.
Another 198 businesses (40%) scored below 40, landing in Not Scored territory. These are businesses that an agent functionally cannot interact with at all. They might have a beautiful website, but if an agent cannot parse it, it might as well not exist.
At the top, just 51 businesses (10%) reached Silver, a single company achieved Gold, and Platinum remains empty. Nobody has cracked 90 or above. The agent economy has no FICO-800 equivalents yet.
The SaaS Leaderboard: Who Scores Highest
These are the top-scoring SaaS products out of our 500-business dataset. Notice that even the best barely crack Silver — and only one has reached Gold. View the full rankings on the leaderboard.
Resend
Vercel
Supabase
Slack
Stripe
GitHub
Resendsits alone at Gold with a score of 75. It is the only SaaS product in our dataset to cross the 75-point threshold. The reason is not that Resend has a better API than Stripe or GitHub — it is that Resend was built API-first with agent-friendly patterns baked into every dimension from day one.
Below Resend, a cluster of Silver-tier titans tells an interesting story: Vercel (70), Supabase (69), Slack (68), Stripe (68), and GitHub (67). These are arguably the most developer-friendly companies in the world. And yet they are 7 to 8 points away from Gold. Why? Because great developer experience does not automatically mean great agent experience.
The Universal Weakness: Onboarding and Pricing
The Pattern We Found in Every Vertical
Across all 500 scans, two dimensions are consistently weak regardless of industry, company size, or API quality:
- D3 Onboarding— Almost no SaaS product allows fully programmatic account creation. Even companies with excellent APIs gate signup behind email verification, CAPTCHA, or interactive onboarding wizards. An agent cannot click “I'm not a robot.”
- D4 Pricing Transparency— Pricing lives on marketing pages as styled HTML, not as structured data. There is no
GET /v1/pricingendpoint. Volume discounts are “contact sales.” An agent comparison-shopping between three CRM platforms has to scrape three different HTML layouts and hope the pricing has not changed since last week.
These two dimensions drag down scores across the entire dataset. A SaaS product can have a perfect API (D2: 95), bulletproof security (D7: 90), and five-nines reliability (D8: 95) and still score Bronze because an agent cannot sign up or figure out how much it costs.
This is the paradox at the heart of SaaS agent readiness: the companies that built the best APIs also built the most human-centric onboarding flows. They invested millions in interactive product tours, beautiful pricing pages with toggle switches, and signup flows with progressive disclosure. All of that is invisible to an agent. The better the human experience, the more likely it is hostile to machines.
Consider Stripe. Its API documentation is a masterclass. Its developer experience scores 90 in our D9 dimension. But ask an agent to create a Stripe account, pick a plan, and start processing payments — programmatically, without a human — and it hits a wall. Stripe requires human-driven signup with identity verification. The onboarding gate drops a world-class API into Silver.
How We Score: The 9 Dimensions
Our v4 scoring engine uses a service-quality-first weighting model. API Quality (D2) carries the highest weight at 15%, followed by Reliability (D8) at 13% and Security (D7) at 12%. Here are all 9 dimensions and what they measure. For the full level-by-level framework, see ARL Levels Explained.
Discoverability
API Quality
Onboarding
Pricing Transparency
Payment
Data Quality
Security
Reliability
Agent Experience
The weighting is deliberate: an agent that can call a clean API (D2), trust that it will respond (D8), and verify the connection is secure (D7) can work around weaker discovery or opaque pricing. But the reverse is not true — transparent pricing and easy discovery mean nothing if the API returns HTML error pages and drops connections.
This is why Tier 1 dimensions (D2, D6, D7, D8, D9) account for 60% of the total score. Tier 2 (D1, D3, D4) accounts for 25%. Tier 3 (D5 plus the agent-native bonus for MCP, A2A, and agent-card.json) accounts for the remaining 15%. The formula rewards fundamentals first and agent-native innovation second.
The Bronze Trap: Why Good Companies Get Stuck
Bronze is not a failing grade. A score of 40-59 means an agent can find your product, partially understand what it does, and maybe call some API endpoints. That puts you ahead of 40% of the market. But Bronze is a trap because it feels like enough.
The 250 Bronze-tier businesses in our dataset share a common profile: strong API quality (D2 averages 55-65), decent security (D7 averages 50-60), but cratering scores in onboarding (D3 typically under 35) and pricing (D4 typically under 30). They built the engine but forgot to build the on-ramp.
Here is why this matters commercially: when an AI agent is comparison-shopping — “find me the best email API for a startup” — it can see all the Bronze products. It can read their API docs. But it cannot sign up, it cannot check pricing, and it cannot complete a transaction. So it skips to the Silver or Gold option where it can actually finish the job. Bronze is visible but not actionable. And in the agent economy, visibility without actionability is worthless.
The math of being stuck: 250 companies scored between 40 and 59. The jump from 59 (top of Bronze) to 60 (bottom of Silver) is just one point. But that one point typically requires fixing the exact dimensions companies find hardest to change: onboarding automation and pricing transparency. It is not a technical gap. It is a strategic one.
What the Top 10% Do Differently
The 51 Silver-tier companies and our single Gold-tier outlier share patterns that the other 90% do not. None of these are revolutionary. They are simply the table stakes that most companies have not shipped yet.
Structured error responses everywhere
Every endpoint returns JSON with typed error codes, not HTML error pages. Even 401 Unauthorized responses include a machine-readable error body. Our scanner gives these APIs 87% of what an authenticated 200 response would score.
Machine-readable discovery artifacts
Published llms.txt, agent-card.json, or comprehensive OpenAPI specs at well-known URLs. An agent does not have to scrape documentation pages to understand what the product does.
Programmatic sandbox or test mode
Even if production accounts require KYC, Silver-tier companies offer API-driven test account creation. An agent can start evaluating the product without a human clicking through an onboarding wizard.
Some form of pricing structure
Published pricing in Schema.org markup, a dedicated pricing API, or at minimum well-structured HTML with consistent CSS selectors. Not perfect, but parseable. The bar is low — most companies do not clear it.
Consistent, versioned API design
Proper pagination, consistent field naming, ISO timestamps, and API versioning headers. These are developer experience basics, but they matter even more when the developer is a machine that cannot guess at inconsistencies.
Rate limit headers and status pages
Silver-tier products tell agents how fast they can call the API and whether the service is currently healthy. These signals let agents make real-time decisions about which services to use.
The pattern is clear: Silver-tier companies do not have fundamentally better APIs than Bronze-tier ones. They have more completeAPIs. They close the gaps in the dimensions that most companies ignore. They publish the metadata that makes agents self-sufficient. The difference between the 50th percentile and the 90th percentile is not brilliance — it is coverage.
The Path from Bronze to Gold
Based on what we see in the data, here is the highest-ROI roadmap for a typical Bronze-tier SaaS product to reach Gold. These are ordered by impact-per-effort.
Publish agent-card.json and llms.txt
A static JSON file at /.well-known/agent-card.json and a plain-text llms.txt at the site root. Zero code changes, zero risk. This is the single highest-ROI improvement any SaaS company can make for agent readiness today.
Add a pricing API or structured pricing markup
Either a dedicated /v1/pricing endpoint returning JSON with plan names, prices, and feature lists, or Schema.org Product/Offer markup on the existing pricing page. Agents need to comparison-shop, and they cannot parse your beautifully designed pricing toggle.
Enable programmatic sandbox creation
An API endpoint that creates a test or sandbox account and returns API credentials. Even if production accounts require human verification, a sandbox lets agents evaluate your product. This is the gate that keeps Stripe, GitHub, and Slack stuck in Silver.
Ensure all error responses are structured JSON
Audit every endpoint — including auth failures, 404s, and 500s — to return JSON with a machine-readable error type, a human-readable message, and a documentation URL. A structured 401 is worth 87% of a 200 in our scoring engine. An HTML error page is worth nearly zero.
Combined impact: A Bronze-tier SaaS product that ships all four of these changes could see a 30-55 point improvement across the weighted dimensions. That is enough to jump from Bronze (43) past Silver (60) and into Gold (75) territory. The work is not hard. The hard part is recognizing that agent readiness is not just an API problem.
Frequently Asked Questions
What is an Agent Readiness Score?
An Agent Readiness Score measures how well a business can be discovered, understood, and transacted with by AI agents. It is scored 0-100 across 9 dimensions including API quality, security, onboarding, and pricing transparency. Think of it as a FICO score for the agent economy — the higher you score, the more likely agents are to choose your product over a competitor.
Why does the average SaaS company only score 43?
Most SaaS companies were built for human users, not AI agents. They have login-gated dashboards, pricing pages that require clicking through marketing copy, and onboarding flows that assume a human is clicking buttons. An excellent API is necessary but not sufficient — agents also need machine-readable discovery, transparent pricing, and programmatic onboarding. These are the dimensions where even well-built SaaS products collapse.
What is the difference between Bronze and Silver tier?
Bronze (40-59) means agents can find your product and partially interact with it, but significant friction points block a full agent workflow. Silver (60-74) means an authenticated agent can complete most tasks end-to-end. The jump from Bronze to Silver typically requires fixing onboarding (D3) and pricing transparency (D4) — the two universal weak dimensions across SaaS.
How can my SaaS company improve its score?
Start with the highest-ROI changes: publish an agent-card.json file and llms.txt for discovery (D1), expose pricing via a structured API endpoint or Schema.org markup (D4), and add a programmatic signup or sandbox creation flow (D3). These three changes alone can move most SaaS products from Bronze into Silver. Run a free scan at agenthermes.ai/audit to see your exact dimension breakdown.
Why is Resend the only Gold-tier SaaS product?
Resend scores 75 because it was designed API-first from day one. Its documentation is machine-readable, pricing is transparent and structured, onboarding is fully programmatic (create an account and send your first email via API in under 60 seconds), and its error responses are clean JSON with typed error codes. Resend does not have a better API than Stripe — it has a better agent experience across all 9 dimensions simultaneously.
Score your SaaS business free
See where you land across all 9 dimensions. Find out if you are Bronze, Silver, or the rare Gold. Get a specific roadmap to level up.