The Investor's Guide to Agent Readiness: How to Evaluate a Company's AI Infrastructure
Every technology investment thesis eventually includes a platform shift question: will this company win or lose when the next wave hits? The agent economy is that wave. We scanned 500 businesses and found that the average Agent Readiness Score is 43 out of 100. The gap between Silver-tier companies and the rest maps directly to AI defensibility. Here is how to add agent readiness to your due diligence framework.
The Investment Thesis: Agent Readiness = AI Defensibility
AI agents are becoming the primary interface between consumers and businesses. When someone asks Claude to “find me the best project management tool for a 10-person team,” the agent queries APIs, reads documentation, evaluates pricing structures, and makes a recommendation. The business with the best API, the clearest documentation, and the most agent-accessible infrastructure wins that recommendation.
This is not speculative. Stripe processes agent-initiated API calls today. GitHub Copilot integrates directly with the GitHub API. Shopify apps increasingly use agent-driven interactions for store management. The businesses scoring Silver and above on agent readiness are already capturing AI-driven traffic. The businesses scoring below 40 are invisible to this channel entirely.
For investors, agent readiness is a leading indicator of technical quality and strategic positioning. A company with high D2 (API Quality), D7 (Security), and D8 (Reliability) scores has built infrastructure that is both agent-ready and generally excellent. These three dimensions represent 40% of the total score weight and correlate with the technical moats that drive long-term defensibility.
The Agent Readiness Due Diligence Checklist
Six metrics to evaluate in every technology investment. Each maps to specific dimensions in the AgentHermes scoring model and can be assessed with a free 60-second scan.
Agent Readiness Score
The single metric that captures API quality, security, reliability, documentation, pricing transparency, and agent experience across 9 weighted dimensions. Run a free scan at agenthermes.ai/audit.
Threshold: Silver (60+) = defensible. Bronze (40-59) = catching up. Below 40 = tech debt liability.
MCP Server Presence
Does the company have a Model Context Protocol server? MCP is the emerging standard for how AI agents discover and interact with services. Presence indicates forward-looking infrastructure investment.
Threshold: Has MCP = ahead of 99% of businesses. No MCP = standard, but needs a roadmap.
API-First Architecture
Is the product built API-first or web-first? API-first companies (Stripe, Twilio, Plaid) are inherently more agent-ready than web-first companies that bolt on APIs later.
Threshold: Documented public API with versioning = strong. No API or internal-only API = structural risk.
Developer Documentation Quality
AI agents consume documentation to understand what an API does. OpenAPI specs, interactive playgrounds, and structured error codes directly improve agent interaction success rates.
Threshold: OpenAPI spec + examples + error reference = Silver-tier docs. PDF-only or wiki = liability.
Security Posture (D7)
TLS everywhere, OAuth 2.0 or API key authentication, rate limiting with transparent headers, and structured error responses. Security-first APIs are also agent-first APIs.
Threshold: D7 score above 70 = production-grade. D7 below 50 = vulnerabilities agents will expose.
Discovery Infrastructure
Does the company have agent-card.json, llms.txt, robots.txt configured for AI crawlers, and structured data markup? These are the signals AI agents and search platforms use for discovery.
Threshold: All discovery files present = proactive. Zero = invisible to AI-driven channels.
Reading the Score: What Each Dimension Tells an Investor
The Agent Readiness Score is composed of 9 weighted dimensions. Here is what the six highest-weight dimensions signal about acompany's technical infrastructure and competitive position.
The moat formula: Companies with high D2 + D7 + D8 scores (combined weight 0.40) have a sustainable technical moat. These dimensions measure API architecture, security posture, and operational reliability — the hardest things for competitors to replicate. High scores here predict that the company will capture agent-driven revenue as the channel grows, because agents prefer reliable, secure, well-structured APIs. See the ROI calculator for revenue impact modeling.
Portfolio Screening: Tier-by-Tier Framework
Scan every company in your portfolio. Rank by score. Then use this framework to categorize risk and opportunity.
Examples: No company has reached Platinum yet
Investor view: Category-defining agent infrastructure. First-mover advantage in the agent economy. Premium valuation justified by defensible moat.
Examples: No company has reached Gold yet (Supabase and Vercel at 69 are closest)
Investor view: Exceptional technical infrastructure. Strong API, documentation, and security. Ready for agent-driven revenue with minor additions (MCP, agent-card).
Examples: Stripe 68, GitHub 65, Slack 68, Vercel 69, Supabase 69
Investor view: Production-grade API with good documentation and security. Can capture agent traffic today. Defend-and-extend position.
Examples: Shopify 52, HubSpot 48, Salesforce 45, Most SaaS companies
Investor view: API exists but has gaps. Agents can interact with friction. 6-12 month runway to reach Silver. Investment in agent readiness needed.
Examples: Most local businesses, legacy enterprises, government services
Investor view: No meaningful agent infrastructure. High risk of disintermediation by agent-ready competitors. Tech debt liability unless addressed.
The Disintermediation Risk
Companies with low agent readiness scores face a specific, quantifiable risk: disintermediation by agent-ready competitors. When AI agents compare options for users, they interact with APIs, not websites. A company with a clean API, structured pricing data, and agent discovery files will be recommended. A company with a PDF price list and a “contact us” form will be skipped.
This is already happening in developer tools. When a developer asks an AI assistant for a payment processing recommendation, the assistant can interact with Stripe's API to check pricing, test features, and read documentation. It cannot do the same with a payment processor that has no API. The recommendation goes to Stripe every time. As agent adoption spreads from developer tools to broader SaaS, e-commerce, and services, this pattern will repeat in every vertical.
The agent economy market size is projected to reach $47 billion by 2028. Companies that are invisible to AI agents will miss an increasingly large share of customer acquisition and transaction volume. For portfolio companies in competitive markets, this is not a nice-to-have metric — it is a survival indicator.
The parallel to SEO: In the 2010s, companies that ignored SEO watched competitors capture organic search traffic. The remedy was expensive and time-consuming: years of content creation, link building, and technical optimization. Agent readiness is the same dynamic at an earlier stage. Building now is 10x cheaper than catching up later. See our CTO guide for the technical implementation roadmap.
How to Scan Your Portfolio
Running an agent readiness assessment across your portfolio takes minutes, not months. Here is the process.
Scan each company at agenthermes.ai/audit
Enter the domain for each portfolio company. Each scan takes 60 seconds and evaluates all 9 dimensions. No authentication or setup required.
Export scores to a comparison spreadsheet
Record the overall score, tier, and individual dimension scores for each company. Pay special attention to D2, D7, and D8 as they signal core technical quality.
Identify the gap portfolio
Companies scoring below 40 need immediate attention. Companies between 40-59 need a 6-month roadmap to Silver. Companies at 60+ are positioned and need maintenance.
Include in board-level reporting
Agent readiness scores belong alongside NPS, ARR, and churn in quarterly board decks. The score trends over time reveal whether the company is building for the AI future or falling behind.
Frequently Asked Questions
Why should investors care about agent readiness now?
Because the window for establishing agent-ready infrastructure is closing. Companies that build agent-accessible APIs, MCP servers, and discovery files now will capture AI-driven revenue streams as agent adoption grows. Companies that wait will face the same competitive disadvantage as businesses that were late to mobile or late to the web. The data from our 500 scans shows that average agent readiness is 43/100 — the entire market is early, which means the gap between leaders and laggards will widen rapidly.
How does agent readiness correlate with company valuation?
We do not yet have enough longitudinal data to prove direct valuation correlation, but the proxy metrics are strong. Companies with high agent readiness scores tend to have high developer satisfaction (strong API = strong developer experience), lower customer acquisition costs (agents recommend them), and more defensible technical moats. The scoring model weights D2 (API), D7 (Security), and D8 (Reliability) highest — the same dimensions that predict technical quality in traditional engineering due diligence.
Can a company improve its agent readiness score quickly?
Yes. The fastest improvements come from discovery infrastructure: adding agent-card.json, llms.txt, and structured data can add 10-15 points in a week. Documenting rate limits and exposing rate-limit headers adds another 5-8 points. These are configuration changes, not architecture changes. Deeper improvements — API versioning, structured error responses, MCP servers — take 3-6 months but yield 20-30 points. We have documented the full improvement path in our Bronze-to-Silver and Silver-to-Gold guides.
What is the biggest red flag for an investor evaluating agent readiness?
No public API at all. A company with no API in 2026 has a fundamental architecture problem that affects more than agent readiness — it indicates a lack of platform thinking, no developer ecosystem, and no integration strategy. The second biggest red flag is a D7 Security score below 40, which means the company has significant security gaps that agent traffic will amplify. Agents probe APIs systematically, and they will find vulnerabilities that casual human users miss.
Should I scan my entire portfolio?
Yes. AgentHermes scans are free at agenthermes.ai/audit and take 60 seconds each. We recommend scanning every portfolio company, ranking them by score, and then categorizing: Silver+ companies are positioned for the agent economy, Bronze companies need a 6-month roadmap, and sub-40 companies need immediate architectural assessment. This gives you a quantified view of AI infrastructure risk across your portfolio.
Scan your portfolio companies for free
Enter any domain and get an Agent Readiness Score in 60 seconds. See exactly where each company stands across all 9 dimensions.