# Brief: agent readiness audit and fixes for [your site]

**From:** [name, role]   **To:** [agency or developer]   **Date:** [date]

## Goal

AI agents (assistants such as ChatGPT, Claude and Perplexity, and browser agents) can reach, read and use our product pages, cart and checkout, so they can recommend us and help shoppers buy.

## Background

- Baseline AgentScore report dated [date], attached. Score [score] out of 100. Main findings: [list].
- Our decision on AI crawlers: [allow AI assistants and AI search; training crawlers: allow / block].
- Platform: [platform name]. Answers from our platform and CDN vendor, where we have them: [attached / to follow].

## In scope

- **Templates:** home, category or collection, product, cart and checkout (where our platform lets you change it)
- **Site files:** robots.txt, sitemap, llms.txt
- **Shared components:** header, navigation, search, pop-ups and cookie banner, variant picker, add-to-cart button
- **Settings:** guest checkout, CAPTCHA settings, [other platform admin settings]

## Out of scope, and who owns it

- CDN and bot protection rules: [security team / CDN vendor, contact]
- Payment provider settings: [finance, contact]
- [Other area]: [owner]

If a finding sits in an area you can't change, raise it with the owner above and record it in your reporting. Please don't park it.

## Deliverables

1. **Audit.** A review of the baseline report that assigns every finding to you or to a named owner.
2. **Fixes.** Changes to the templates, files, components and settings in scope.
3. **Change list.** For each fix: what changed, the pages and components it touched, and the check or journey it was for.
4. **Blockers.** What couldn't be fixed and why, with the owner it was handed to and the date.
5. **Final results.** A final scan report of the live site, and the results of each named journey test.
6. **Upkeep note.** What to re-check after future releases, such as theme updates, new apps or new pop-ups.

## Journeys to test

- [Find a product and reach checkout]
- [Use search to find a product in a given size]

Test with an AI agent, using test data, and stop before payment. Run each journey several times, because AI agents vary.

## How we'll measure success (acceptance criteria)

1. These checks pass in a scan of the live site: [list AgentScore check titles, such as "Product pages give price and stock in a form agents can read"].
2. No check that passed in the baseline scan now warns or fails.
3. Each named journey passes in every run, not just some.
4. Any check marked "not tested" comes with the scan's reason and a manual test.

We'll judge the work by these checks and journeys, not by a target score.

## Timeline

- **Week 1:** Baseline. Scan, journeys and brief agreed.
- **Weeks 2 to 4:** Fixes. Change list updated as work ships.
- **Week 5:** Acceptance. Final scan and journeys against the criteria.
- **After:** Upkeep. Re-check after each release.

## Contacts

- Our owner: [name]
- Security or CDN: [name]
- Platform: [name]
