Apply a fix plan and check it worked

Steps

1. Run an audit

Add your first site and AuditAE runs the workspace's first audit automatically: 4 buyer questions on ChatGPT, Perplexity, Gemini and Google AI Overviews. After that, run one from the site's Overview (Run check now) or Citations → Audits (New audit). See Add a site and run an audit.

2. Read the diagnosis

Every audit ends with a short list of top fixes (1–3 actions ranked by how many lost answers each would close), and every answer that doesn't cite you gets a playbook row. There are two kinds of loss: mentioned-only (the engine named your brand but cited someone else's URL, the most fixable kind) and pure loss (your brand wasn't mentioned at all). See Read audit results.

3. Get the exact edits

Ask for the sentences to rewrite, the FAQs, title and meta, JSON-LD and missing pages for a page or the whole site. In the dashboard, each top fix has an AEBOT button: Apply with AEBOT when the site is a paired WordPress site with Write SEO settings on, Show me the exact edits when it's paired without it, and Do it with AEBOT otherwise. A reply is usually about $1–2. Over MCP, call get_fix_plan ($0.50 per plan; the same request again within 24 hours, errors, and plans with nothing left to fix are free). See What's in a fix plan below.

4. Apply the edits

Pick the route that matches the site:

• Paired WordPress site: AEBOT or your MCP client applies the fixes with the wp_* tools. Post edits need Create + update draft posts or Publish posts directly; SEO titles, meta, schema, FAQ schema and llms.txt need Write SEO settings. Each write costs 5¢. See What AEBOT can do at each WordPress permission level.
• Shopify, Webflow, Wix, Squarespace, Framer or another builder: follow each fix's apply steps for that platform and paste its ready-made code (JSON-LD, meta tags) where it says.
• A site built from code (Next.js, Astro, …): have Claude Code or Cursor, connected to AuditAE, edit the repo from the plan. See Connect Claude, ChatGPT or Cursor.

Never publish a fix marked as a draft until you've confirmed or supplied the facts it lists.

5. Tell the crawlers the pages changed

Once the changes are live, call submit_urls with the changed URLs (free). It pings IndexNow, which notifies Bing, Yandex, Naver and Seznam so they re-crawl sooner. Google doesn't use IndexNow. Bing's index is one of the sources ChatGPT search and Copilot draw on. On a paired WordPress site the AuditAE plugin already pings IndexNow when you publish or update, unless you turned its auto-ping off.

6. Check it worked

Engines take days to weeks to pick up a change. Over MCP, schedule_recheck re-asks the plan's prompts N days later against the audit you started from (free to schedule; the checks and their playbook are billed when it runs), and get_recheck compares citations before and after. In the dashboard, the first audit comes with a free re-check: 14 days later we re-ask up to 5 of its questions on ChatGPT and Perplexity and email you whether you moved. To keep watching, use the Re-check these N questions every week card on your first audit's results and click Track weekly.

What's in a fix plan

A fix plan is a ranked list of exact changes that get your brand cited by AI engines instead of competitors. It's written for the platform the site runs on, and it checks the live site first (sitemap, llms.txt, robots.txt, homepage links, JSON-LD). Each fix carries:

The plan also returns the prompts to re-ask once you've shipped (verify_prompts) and the audit it was built from, which schedule_recheck uses as the baseline.

Fixing the share-of-voice gap

When buyers see your brand name but click a comparison-site URL, that's an attribution leak: the engine answered about you, but a third party got the credit. Fix it with on-page schema, comparison content, or partnerships with the sites that are getting cited.

Every audit ends with a "do this first" list: 1–3 actions ranked by how many losses each would close. If the site has a brief in Memory (goals, ideal customers, decisions already made), the ranking respects it: it won't re-propose something you've marked as tried, and it favors moves that serve your stated priority keywords.

What it costs

• The playbook and top fixes on your first audit aren't charged — only the engine checks use credit.
• Audit checks: priced by engine per prompt × engine cell: Perplexity 5¢, ChatGPT 7¢, Google AI Overviews 9¢, Gemini 12¢.
• Playbook rows: one per answer that doesn't cite you, about 10¢ each, plus about 45¢ for the top-fixes summary.
• get_fix_plan: $0.50 per plan. Repeats within 24 hours and errors are free.
• AEBOT: billed on the reply's tokens; a reply is usually about $1–2, and AEBOT needs $1.50 of credit to start one.
• WordPress writes: 5¢ each. Reads are free.
• Re-checks: free to schedule. When one runs, each check is billed at the engine's price, plus playbook rows for answers that still don't cite you.

Full price list: Credits, prices and top-ups.

Any website, not just WordPress

get_fix_plan detects what the site is built on and returns each fix with steps for that platform plus paste-ready JSON-LD and meta tags. On a paired WordPress site the agent applies it with the wp_* tools; on a code-built site (Next.js, Astro, …) a coding agent like Claude Code or Cursor edits the repo; on Shopify, Webflow, Wix, Squarespace or Framer it walks you through the editor.

Further reading

• How to rank on ChatGPT: the strategy behind these fixes.
• Pair the AuditAE WordPress plugin so AEBOT can apply fixes for you.