--- url: https://ora.ai/blog/what-agents-actually-reach title: "What agents actually reach" description: "Real agents lean hard on the human web - homepages and docs, in the same order a person would - and barely reach the files we built for them. Where those files do get reached, they convert better than anything. The bottleneck is discovery." crawled_at: 2026-10-18T19:27:07Z lastmod: 2026-06-16 ask: https://ora-ai.ora.run/ask docs: https://ora-ai.ora.run/.well-known/ora/docs lang: en --- # What agents actually reach ora research · Jun 16, 2026 · 5 min read [Share on X](https://x.com/intent/tweet?text=What%20agents%20actually%20reach&url=https%3A%2F%2Fora.ai%2Fblog%2Fwhat-agents-actually-reach) [Share on LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fora.ai%2Fblog%2Fwhat-agents-actually-reach) We watched real agents work - actual episodes of agents landing on websites and trying to get things done, drawn from real runs rather than a benchmark built around the formats we ship. One question: what do they actually reach, in what order, and what do they do with it? The answer is a little uncomfortable for anyone who’s spent the last year shipping llms.txt files. Agents lean hard on the human web. They reach for homepages and docs - the same pages a person would, in the same order. The purpose-built, agent-first files get used far less. But where those files exist and get reached , they convert better than anything else. The bottleneck is discovery. Here’s the evidence. finding 01 ## Everything that isn’t docs falls off a cliff How often each resource got used, once it existed on a site: Usage rate when present all sites docs pages · 88 % · homepage · 84 % · the cliff · llms.txt · 46 % · .well-known/* · 34 % · openapi.json · 17 % · robots.txt / sitemap.xml · 9 % · agents.md · 8 % · llms-full.txt · 5 % Usage rate when present, across all sites. Docs and homepages live in one tier. Everything we built specifically for machines lives in another, far below. The drop from documentation to the rest is a cliff. finding 02 ## But agent-first formats win when they get a shot On the one site in the set that shipped the full stack - and linked it properly - the numbers invert: The inversion all sites → full-stack site openapi.json + 55 pts when reached all sites · 17 % · full-stack site · 72 % .well-known/* + 29 pts when reached all sites · 34 % · full-stack site · 63 % agents.md + 34 pts when reached all sites · 8 % · full-stack site · 42 % Usage rate when reached: across all sites, versus the one site that ships the full stack and links it. Same files. ~17% → ~72% . The difference is reachability. Reachability is the difference. TRACE ora.ai ships the full stack and links it - so the agent reaches openapi, .well-known/*, agents.md and llms-full.txt in a single clean run. 17 steps, 1 search. finding 03 ## Agents search to find their way, not the answer Most homepage and docs visits came from prior brand knowledge - the agent already knew where it was going. Web search showed up in ~38% of runs. But when it did, it mostly handed the agent a URL , not a fact. The task still got completed from on-site content. Search is the map. The site is the destination. TRACE attio.com, finding pricing: 38 steps, 6 searches. Search keeps pointing the agent at the next URL - docs and help pages - while the answer gets assembled from what’s on-site. finding 04 ## AI-native files are reached late, and only by following links Order of arrival, by turn: Order of arrival earlier → later turn 1.0 · homepage · turn 1.7 · docs · turn 3.0 · llms.txt / openapi.json · turn 4.6 · agents.md · turn 6.7 · llms-full.txt Familiar pages first. Structured files many turns later - and only if something linked to them. A file nothing points to gets reached late, or never. An llms.txt no page references is a dead end. finding 05 ## The homepage is the gateway, and the easiest place to break a run Reached in ~84% of runs , almost always on turn one (~95%) . It’s the front door for nearly every session. Which makes it the most dangerous place to fail. Hide the navigation behind JavaScript and the agent goes blind from step one. Everything downstream depends on clearing this gate. finding 06 ## Agents guess standard paths, so meet them where they look Agents probe conventional URLs from habit. /pricing and /integrations get hit early, and they work when they exist. But habit runs ahead of reality: /api - guessed in ~11% of runs, found in ~2% . Agents expect a convention most sites don’t serve. The fix is the cheapest one on this list: serve the paths agents already reach for. TRACE telnyx.com, developer finding API docs: the run succeeds via the llms.txt branch - but watch the orange node. The agent probes .well-known/api-catalog on spec; when the convention is missing, that guess is a wasted reach. finding 07 ## Whatever an agent reaches, it uses, so wrong content is worse than none Once a file is fetched, it shapes the answer almost every time. Agents build from the first believable page they land on. That cuts both ways: + Good content gets used, readily. ! A reachable page with stale or wrong info hurts more than no page at all, because the agent will trust it and run. Accuracy is load-bearing. the one takeaway ## Discovery is the bottleneck Agents behave like well-informed visitors: they arrive at the front door from memory, fan out to docs, and follow links from there. The files we build for them are genuinely effective. They lose on discovery. So the work is making the machine-readable formats reachable : - Link to them from the pages agents already hit. - Serve the conventional paths agents already guess - /api , /pricing , /integrations . - Keep the homepage navigable without JavaScript. - Treat every reachable page as gospel - because to an agent, it is. Meet agents where they already look, and the formats built for them finally get their shot. ### Want to see where you rank? Run the same scan we ran on thousands of sites. Free, public, takes about 1 minute. Scan your site → Explore the data ← all posts # Footer The standard for agent experience. © 2026 [era labs](https://eralabs.ai/) Product - Scan a site - [Journey](https://journey.ora.ai/) - [WebMCP audit](https://webmcp.ora.ai/) - [Directory](https://directory.ora.ai/) Resources - Methodology - Research - Blog - API Docs Company - About - Contact - Privacy - [@oradotai](https://x.com/oradotai)