See where your product stands against competitors on AI discoverability and agent usability.

Today, we’re launching Agent Experience Benchmark by The Prompting Company.

It ranks software, infrastructure, and developer tool companies by how well AI agents can discover, understand, and use them.

The question is simple: when an agent is choosing between products, where does yours stand?

Agents are becoming product evaluators

AI agents are starting to do more than answer questions.

They search for products, compare vendors, read docs, choose APIs, install packages, write code, and attempt workflows.

That changes what it means to win a customer.

In the old journey, a buyer might search Google, read a few pages, talk to sales, and decide which product to try.

In the new journey, an agent may do the first pass. It may decide which products are worth considering, which docs to trust, which package to install, and whether the product works well enough to continue.

Software companies need a way to measure that new journey against competitors.

That is what Agent Experience Benchmark does.

What the benchmark measures

For each category, we score how well an agent can move through two parts of the journey: discovery and usability.

Discoverability measures whether an agent can find, read, and recommend your product when it is looking for a solution in your category. We look at whether your product shows up in relevant AI answers, how often it is mentioned, how it ranks against competitors, which sources are cited, and whether your website gives agents enough context to understand what your product does.

Usability measures whether an agent can actually build with your product once it finds it. We run live coding agents through product-specific workflows and score whether they can find the correct docs, resolve dependencies, set up the environment and authentication, avoid deprecated paths, execute the task, and verify that it worked.

Together, these scores create a composite view of Agent Experience.

How the leaderboard works

The leaderboard ranks products within the same category.

Each product gets an Agent Experience score, grade, and ranking against competitors. You can also see the category it belongs to and when the score was last updated.

The leaderboard is designed to make comparison easy. Instead of looking at products one by one, teams can see where they stand in their category and which competitors are ahead or behind.

Score your own product

If your product is not on the leaderboard, you can click “Score your product” to run an audit.

The audit shows how your product performs across the same journey measured in the benchmark, with a deeper breakdown of where agents succeed or get stuck.

You’ll see how your brand appears across AI engines, how you rank against competitors, which models and topics mention you, and whether your website is ready for agents to discover and read.

You’ll also see whether agents can use your product to complete real workflows, including your usability score, pass rate, failure clusters, and action items your team can use to improve the experience.

The goal is to give every company the same view the leaderboard provides: where they stand, why they scored that way, and what to improve.

Why this matters now

Software companies have spent decades optimizing websites, docs, and onboarding for humans.

Now agents are becoming part of that journey, deciding which products to consider, which docs to trust, and whether a product works well enough to keep using.

The companies that win the next decade will not just be discoverable by AI. They will be usable by agents.

Agent Experience Benchmark is our first step toward making that measurable.

We’re starting with software, infrastructure, and developer tools, where agent workflows already happen through docs, APIs, SDKs, CLIs, and MCP servers.

If your next user is an AI agent, your product needs to be built for agents too.