For AI agents: a documentation index is available at /llms.txt
AXO · Readable and operable
Agent-ready websites
WorkPods makes your site agent-ready across the Readable and Operable layers of AXO (agent experience optimization). A site is agent-ready when an AI assistant can fetch your pages without running scripts and use your forms without getting stuck on unlabelled controls.
# /llms.txt
# Example Co
> Invoice automation for finance teams.
## Pricing
- [Plans](https://example.com/pricing.md)
Before<div class="field">Company size</div>
After<form toolname="book_demo"> <label for="size">
01
Make your pages readable and operable for agents
Sites built for people in browsers often trip agents up for two reasons: key copy that appears only after JavaScript runs, and demo forms with unlabelled inputs. Either one can stop an agent from reading your offer or booking a meeting.
Free checkers such as Cloudflare's isitagentready.com and Chrome's agent-ready toolkit give useful surface context. WorkPods has no partnership with either company, and a surface score will not repair a form an agent cannot fill in. Our own free extension, Agent Ready, runs the page checks from our audits in your browser and points to the element behind each failure. Your pages also need crawlable HTML that fits your technical SEO, plus structured data (labels telling machines what a page is). The forms that convert visitors have to be clear enough for an automated browser to operate safely.
02
Run a 10-minute self-test on your website
You can check whether your site is agent-ready in ten minutes, with no code and no wait for engineering. The four steps below show where AI tools lose your content or drop out of your conversion flow.
| Self-test step | How to run it in plain steps | What a pass looks like | What a fail looks like |
|---|---|---|---|
| 1. Find pricing and summarise your offer | Ask an assistant to visit your domain, summarise what you sell and list your pricing tiers | The assistant states your category, target buyer and plan structure accurately from your pages | It guesses from old third-party mentions or says it cannot read your pricing page |
| 2. Walk through booking a demo | Ask a browser-capable assistant to open your demo page and list the fields needed to request a call (stop before submitting) | It locates the form and reads every input label and dropdown option clearly | It cannot find the booking button or gets stuck on unlabelled fields and custom dropdowns |
| 3. Check pages with scripts turned off | Turn off JavaScript in your browser settings and reload your homepage, top product page and pricing page | All core headlines, body copy, plan tables and FAQs still appear on the screen | The page shows a blank screen, a loading spinner or missing pricing tables |
| 4. Check your robots file and site guide | Open yourdomain.com/robots.txt (the file setting bot rules) and yourdomain.com/llms.txt (a proposal at llmstxt.org telling AI tools what your site covers) | Your robots.txt sets clear rules for AI crawlers and Content-Signal, and llms.txt links to key pages | Your robots.txt blocks all AI bots by default and llms.txt returns a broken file |
If step 3 or step 4 fails, start there. Agents that cannot read a page never reach the form on it.
03
Let agents read your files and finish key tasks
WorkPods hands over findings with evidence, plus ready-to-use files and code patches your web or CMS team can deploy straight away. The files help answer engines read your site, and the form fixes help browser agents finish tasks on it.
On the Readable side, we draft your llms.txt file, Markdown versions of your pages, and Content-Signal and AI-crawler rules in robots.txt. We also prepare JSON-LD structured data patches and HTML changes so your copy loads without JavaScript.
On the Operable side, we fix form field labels and add WebMCP markup to key forms. WebMCP is an emerging standard that lets a page declare tools for browser agents. AI agents run the first checks and drafts, an operator reviews every one, and your team approves each change in our approval log. The full loop is on how WorkPods works.
04
How we measure it
We measure your site before and after each fix, using defined file checks and safe browser-agent runs on your money pages. Each result has a fixed status, so your marketing and web teams can see which blockers are gone and which remain.
- Each AXO check is scored Pass, Partial, Fail or Not applicable. A missing optional file like llms.txt is Not applicable, never a fail, and only a present, broken file gets Fail.
- We test 3 to 5 key tasks from your money pages confirmed by your team: find pricing, book a demo, find a contact. Each task ends Completed, Reached the confirmation step, Partial or Blocked, recorded with the step where it stopped.
Our test agent never submits a form or buys anything on your site, and nothing reaches your CMS without your approval. To see which bots actually fetch your pages after the fixes, pair this work with agent analytics. To see where you stand today, request a free GEO and AXO audit.
Questions
Questions buyers ask
How do I know if my website is blocking AI agents?
Check your robots.txt file and your server logs to see whether AI crawlers and user-fetch bots are blocked. Security tools and CDNs can block automated requests, which stops assistants from reading your pricing or product pages. Our audit reviews your public bot rules and tests whether agents can fetch your core pages cleanly.
What is llms.txt and does a missing file count as a fail?
An llms.txt file is a short Markdown guide at your site root, based on the proposal at llmstxt.org, that tells AI tools which pages explain your product and pricing. A missing llms.txt counts as Not applicable, never a fail. Only a file that is present and broken gets Fail, and we draft a clean one for you.
Why does JavaScript stop AI agents from reading our pages?
AI crawlers and assistant fetchers generally read only the HTML your server sends and do not run browser JavaScript. If your product descriptions, comparison tables or pricing tiers load through client-side scripts, the fetcher sees an empty template. Serving core copy in the HTML fixes this for AI agents and search crawlers alike.
What is WebMCP and how does it help conversion forms?
WebMCP is an emerging standard that lets a webpage declare tools for browser agents on key forms. Together with clear HTML field labels, it tells a browser agent how to use your demo request or contact form. That makes it more likely a task ends Completed or Reached the confirmation step rather than Blocked.
Will your testing agent submit fake leads into our CRM?
No. Our browser agent never submits a form or buys anything on your site. We test 3 to 5 tasks your team confirms from your money pages and score each one Completed, Reached the confirmation step, Partial or Blocked. Each result names the step where the agent stopped, and your CRM never sees a test lead.
Who puts these fixes live on our website?
Your own web, CMS or marketing operations team deploys the approved changes, using the files and patches we hand over. WorkPods prepares the draft llms.txt and robots.txt rules, JSON-LD patches, label fixes and WebMCP markup. Nothing ships until you approve it in our approval log, and we retest before and after each fix.