Measuring Your Site's Agent Readiness: The Metrics That Matter

Most businesses measure website performance through a human lens: page load speed, bounce rate, conversion rate, session duration. These metrics tell you nothing about how well your site serves AI agents.

Agent readiness requires its own measurement framework. The good news is that most of the data you need is already available in your server logs and existing tools. You just need to know what to look for.

Tier 1: Visibility Metrics

These tell you whether agents can find and read your content at all.

Server-Side Content Ratio. Load your key pages with JavaScript disabled. Calculate what percentage of your meaningful content (product data, prices, descriptions, reviews) is present in the raw HTML response versus what requires JavaScript to render.

Target: at least 90% of critical content should be in the initial HTML. Many modern single-page applications score below 10% on this metric, which means agents see almost nothing.

Structured Data Coverage. For e-commerce sites, measure the percentage of your products that have complete, valid Schema.org markup. "Complete" means not just a Product type, but also Offer (with price, currency, availability), AggregateRating, Brand, and relevant product attributes.

A coverage score below 80% means agents are making decisions about your products with incomplete information.

Structured Data Accuracy. Coverage alone isn't enough. Measure how often your structured data matches what's actually shown on the page. If your Schema.org markup says a product costs £49.99 but the page displays £54.99, agents that check both sources will flag the inconsistency.

Run automated checks weekly. Accuracy should be above 99%. Anything lower suggests a systemic problem with how your structured data is generated.

Tier 2: Accessibility Metrics

These measure whether agents can successfully interact with your site.

JavaScript-Free Navigation Depth. How many pages can an agent reach by following links in the raw HTML, starting from your homepage? Map the site as a crawler would, without executing JavaScript. Compare this to your full sitemap.

If your JavaScript-free navigation only reaches 30% of your pages, agents are locked out of 70% of your content.

API Response Consistency. If you have public-facing APIs, measure how closely the data they return matches your website content. Check product names, prices, availability, and descriptions. Log any discrepancies with timestamps.

Inconsistencies above 2% indicate data pipeline problems that will affect agent trust.

Critical Path Completion Rate. Define the most important user journeys on your site (finding a product, checking price, adding to basket, completing checkout). Test each path without JavaScript and note where it breaks. Express this as a percentage of completable steps.

Most e-commerce sites score between 20% and 40% on this metric, meaning agents can start but not finish the tasks that generate revenue. The hidden revenue cost of these failures adds up quickly.

Tier 3: Traffic and Engagement Metrics

These show you how agents are actually interacting with your site today.

Agent Traffic Share. Parse your server logs to identify requests from known AI agent user-agent strings. Major agent platforms like those from OpenAI, Anthropic, Perplexity, Google, and Microsoft have identifiable signatures. Calculate agent requests as a percentage of total requests.

Track this monthly. If it's growing (and it almost certainly is), you want to know the rate.

Agent-Specific Error Rate. Filter your server error logs to show only agent-originated requests. Look at 4xx and 5xx error rates specifically for this traffic. Compare to your human visitor error rate.

It's common to find that agent error rates are 3 to 5 times higher than human error rates, usually because agents request pages or resources that require JavaScript to exist.

Agent Session Depth. For agents that make multiple requests in sequence, how many pages do they visit before stopping? A shallow session depth (1 to 2 pages) suggests the agent is failing to find useful content and giving up. A deeper session (5+ pages) suggests productive crawling.

Building a Dashboard

You don't need expensive tools for this. A basic monitoring setup looks like:

Weekly automated tests: Run a headless browser with JavaScript disabled against your top 50 pages, or use Agent Checker to automate the process. Check for content presence, structured data validity, and link accessibility. Store results in a simple database or spreadsheet.

Daily log analysis: Parse server logs for agent user-agent strings. Track volume, requested paths, response codes, and response times. Most log analysis tools can handle this with custom filters.

Monthly structured data audit: Validate your Schema.org markup against actual page content. Flag any mismatches. This can be automated with a simple script that compares JSON-LD data to scraped page content.

Benchmarks by Sector

Based on our analysis of hundreds of sites, here are typical agent readiness scores by sector:

• News and media: Generally strong, around 70 to 85% content visibility, due to their historical investment in server-side rendering.
• E-commerce: Highly variable, from 20% to 75%, depending on platform. Shopify sites tend to score higher than custom-built SPAs.
• Financial services: Often poor, around 25 to 40%, due to heavy reliance on dynamic, authenticated content.
• Travel: Mixed, around 35 to 60%. Booking engines are usually JavaScript-heavy, but informational pages tend to be better.
• B2B SaaS: Typically weak, around 15 to 35%, since marketing sites often prioritise visual impact over machine readability.

These benchmarks give you a starting point. Your actual competitive position depends on where you stand relative to direct competitors, not the industry average.