Skip to content
The Wealth
Experiment
Let’s talk AI
← Agent-friendly websites

Agent-friendly websites

Agent-Friendly Website: A Small Business Test You Can Run

Run a bounded agent-readiness experiment with a baseline, a short llms.txt file, a test log and an honest decision about what to do next.

Wealth Experiment · Published · Updated · 8 min read

Wealth Experiment guide: A decision worth making. Illustrated example, not a measured result.
Original explanatory graphic. The worked scenario is fictional; it is not a client result or performance claim.
In this guide
  1. What should you improve first?
  2. What to build—and what each layer does
  3. Set a budget before opening the editor
  4. Worked example: five questions, one offer
  5. Use three possible decisions
  6. The six-step implementation
  7. A short llms.txt you can adapt
  8. If you need an API, start with a read
  9. A better audit prompt
  10. Test what happened, not what the assistant says happened
  11. Avoid the shortcuts that create new problems
  12. Questions worth answering
  13. Put one improvement to work
  14. Sources and editorial notes
  15. Continue with the right practical resource
  16. Sources and further reading

What should you improve first?

The hypothesis is deliberately modest: clearer public information will make a specific explanation more accurate. It is not “adding llms.txt will create sales.” Keep the technical result separate from demand, traffic and revenue so that a small win does not turn into an unsupported business conclusion.

An evening spent improving a website can be productive. It can also become another way to postpone talking to a customer. This experiment gives the work a boundary: choose one offer, test whether an assistant understands it, fix the largest gap, and decide whether the improvement deserves more time.

The practical answer

The hypothesis is deliberately modest: clearer public information will make a specific explanation more accurate. It is not “adding llms.txt will create sales.” Keep the technical result separate from demand, traffic and revenue so that a small win does not turn into an unsupported business conclusion.

What to build—and what each layer does

LayerUseful jobWhat it does not establish
Readable, accessible pagesExplain the offer and make the next action understandable.That an agent will interpret every detail correctly.
A short llms.txtProvide a concise map for clients that use it.Permission, universal adoption or a search ranking boost.
Documented read-only dataReturn structured facts for a defined task.That those facts are fresh or safe without maintenance.
Optional MCP integrationExpose capabilities to compatible clients.Automatic discovery, authorization or reliable outcomes.

Choose the layer that solves the observed problem. llms.txt proposal and format describes the file convention; OpenAPI specification describes HTTP API contracts; Model Context Protocol architecture explains MCP’s client/server architecture. These are different responsibilities.

Set a budget before opening the editor

Allow one focused session for a baseline, one for the change and one for review. Decide what would make the experiment worthwhile: fewer incorrect answers about scope, a working next-step link or a clearer exclusion. Record setup time as well as test time. Use tools already available to you before subscribing to an audit platform.

Worked example: five questions, one offer

Illustrative Weekend Systems sells a simple planning session. Ask five fixed questions about the audience, deliverable, price basis, exclusions and next step. Save the answers and the source URLs. After editing the offer page and publishing a short map, repeat in fresh sessions. If the second run finds the right price but still invents a guarantee, that is a partial improvement and an unresolved problem—not a successful launch.

Use three possible decisions

Keep the change when it reduces an observed error and is cheap to maintain. Revise it when the answer remains wrong or the source is ambiguous. Stop expanding the experiment when the next step would be an API nobody needs. A new service page still needs customer conversations and distribution; technical clarity cannot establish willingness to pay.

The six-step implementation

Six steps: Write a hypothesis: Pick one observable problem.; Capture the baseline: Save five questions and answers.; Make one change: Clarify the biggest gap.; Repeat the test: Use fresh comparable sessions.; Count the work: Include setup and review.; Choose the next move: Keep, revise or stop.
Use the sequence as a working checklist. Each step produces something you can inspect.
  1. Write a hypothesis. Pick one observable problem.
  2. Capture the baseline. Save five questions and answers.
  3. Make one change. Clarify the biggest gap.
  4. Repeat the test. Use fresh comparable sessions.
  5. Count the work. Include setup and review.
  6. Choose the next move. Keep, revise or stop.

A short llms.txt you can adapt

This is an illustrative template. Replace every example.com URL with an existing public page on your own domain. Check the content, not just the link. If your site has no public API, leave API instructions out. Do not publish placeholder endpoints, private URLs, access tokens or customer information.

# Weekend Systems

> Explain the offer, identify who it is for, and find the correct next step using only the public website.

## Useful pages
- [Main information](https://example.com/offer/): The current offering and relevant details.
- [Conditions and process](https://example.com/questions/): Eligibility, limitations and how the process works.
- [Next step](https://example.com/start/): The official route for a person to continue.

## Boundaries
This example does not support automated purchases or bookings.
Use the linked pages for current facts; ask when information is missing.

Serve the file at your site root when your hosting setup supports it. In a typical static build, that means placing it in the directory copied to the public root. Verify the deployed URL: a local file alone is not a published file. If your platform already generates one, inspect and maintain that version instead of creating conflicting instructions.

If you need an API, start with a read

The following is a proposed contract, not an endpoint installed on this site. Choose fields that answer the task, document their types and meaning, and return only information you intend to make public. Include a source URL and an update signal where useful. A checked date records a review; it does not guarantee the underlying data is current.

GET https://example.com/api/v1/services
Accept: application/json

{
  "example": true,
  "checked_at": "2026-11-11",
  "items": [{
    "id": "intro",
    "name": "Introductory service",
    "pricing_basis": "Request a current quote",
    "source_url": "https://example.com/services/",
    "available_actions": ["read_details"]
  }]
}

Define no-result and error behavior as carefully as success. Keep private data behind proper access controls. For future actions that change records or spend money, design authentication, authorization, explicit confirmation and duplicate-action protection separately. OWASP API Security Project provides a useful security review starting point.

A better audit prompt

Use an assistant with browsing or request capabilities. If it cannot access the site, it should report the limitation; a fluent response is not proof of inspection.

Review this public website: [URL]. Task: Explain the offer, identify who it is for, and find the correct next step using only the public website.

Use browsing or request tools if available; if unavailable, say what you cannot verify. Inspect the relevant public pages and /llms.txt. Record the URLs you actually fetched and the facts you found. Do not infer success from a 200 response alone. Separate observed facts, missing information and suggested changes. Do not sign in, submit forms, change data, add to a cart or purchase.

Draft a concise llms.txt using only verified public URLs. Include an API example only if a documented public endpoint exists and its read operation is appropriate to test. Do not invent endpoints. Rank the three most useful fixes by customer impact and effort. Explain how to retest each one.

Test what happened, not what the assistant says happened

TestEvidence to keepA passing result
Retrieve the mapFinal URL, status, content type and response body.The actual text file appears; no login page or application shell.
Follow its linksThe final destination and the relevant source facts.Useful pages resolve and match the stated purpose.
Try the customer taskExact prompt, tool capability, answer and cited URLs.Required facts and next step are correct.
Ask an unsupported questionThe answer and what the source actually says.The assistant names the limit rather than inventing a capability.
Repeat after a changeSource version, date and all attempted runs.Any improvement is reproducible within the tested setup.

Measure for this site’s use case: Accurate answers out of attempted questions, maintenance time and next-step failures. A small convenience sample is not a causal revenue study.

Run a baseline before editing. Use fresh sessions and keep the prompt and available tools comparable. Include failed attempts in your denominator. Record whether you explicitly told the agent about llms.txt; that is different from observing spontaneous discovery. A scanner is a diagnostic aid, not a certification.

Avoid the shortcuts that create new problems

  • Do not replace readable pages with machine-only content. People still need to evaluate the offer.
  • Do not treat robots.txt or llms.txt as authentication. Protect private routes with real access controls.
  • Do not tell every crawler it may access everything merely to improve a score.
  • Do not duplicate prices and policies into several files without an update process.
  • Do not claim a successful booking, lead or sale from a generated answer. Verify the actual system result.

For interface implementation, consult Google web.dev: agent-friendly website UX and W3C: labeling form controls. Preserve descriptive controls, meaningful form labels and visible state changes. Agent testing complements human usability and accessibility testing.

Questions worth answering

Will llms.txt improve Google rankings?

No ranking benefit is established by adding this file. Google’s current guidance says it does not use llms.txt for Search visibility. Treat the file as an optional usability aid for clients that consume it; keep ordinary SEO work and task testing separate.

Do I need MCP or Markdown copies of every page?

Not for every project. Start with the smallest interface the task requires. Add Markdown resources or a compatible tool integration when there is a clear consumer and maintenance plan. Check platform support rather than assuming all assistants behave the same way.

Can I use this without coding?

Yes. You can clarify public pages, check links and run the task exercise with your existing website editor and assistant. Publishing a root file depends on the host. A public API requires additional implementation; ask for help if that is outside your skills.

How often should I review it?

Review whenever an offering, policy, route or API contract changes. Assign an owner. Frequently changing prices or stock need stronger freshness controls than a static guide.

Search guidance: Google Search: optimizing for generative AI. Documentation reviewed November 11, 2026. Provider behavior can change; recheck the linked sources when implementing.

Put one improvement to work

Start with this task: Explain the offer, identify who it is for, and find the correct next step using only the public website. Save the baseline, make one justified improvement and keep the retest evidence. Use the result to choose the next useful change.

Plan your next experiment →

Sources and editorial notes

This guide combines primary-source documentation with original implementation advice and a fictional worked example. It is not a report of measured client results, guaranteed traffic or earnings. The examples use reserved example.com URLs and must be adapted and verified before deployment.

Sources and further reading

Sources checked November 11, 2026. Review the linked publisher information for current details.