What You Need to Launch AI SEO Yourself

The minimum to launch AI SEO yourself: a working website (CMS or static site), a list of the 20 questions your buyers ask most, one AI writing tool (Claude or ChatGPT Pro), schema markup capability (plugin or AI coding tool), and roughly 30 hours spread over 6 weeks. Everything else (tracking tools, llms.txt, advanced analytics) is optional and can be added after the cluster is live.

The 5 non-negotiables

Without all 5, the launch does not work.

1. A working website

A site that Google and Bing can already crawl and index. Can be any platform:

  • WordPress with a modern theme
  • Webflow, Framer, or similar no-code builders
  • A static site generator (Hugo, 11ty, Astro)
  • Shopify (if AEO is for the ecommerce angle)

What does not work: Google Sites, a Linktree, a single static HTML page with no CMS, anything that hides content behind JavaScript with no server-side rendering.

If you do not have a working website yet, that is step 0. Do not try to do AEO on a broken or non-existent site (developers.google.com).

2. The 20 questions

The cluster stands or falls on question quality. Sources for the 20:

  • Questions you have heard from buyers in sales calls (the best source)
  • Questions your buyers actually type into ChatGPT right now (run them yourself and see what comes up)
  • Questions your competitors already rank for (use Ahrefs, SEMrush, or Google's autocomplete)
  • Questions from your own support inbox or community threads

Combine sources, then cut to 20 that matter most for buying-stage buyers, not trivia-stage curious people.

3. An AI writing tool

Claude Pro, ChatGPT Plus, or equivalent at $20 per month. Free tiers work but the quality delta makes the upgrade worth it for content production work.

Optional upgrade: Claude Code for handling schema generation, internal linking, and file management at the same time. Saves hours per cluster.

4. Schema markup capability

Three ways to add schema (pick one):

  1. CMS plugin. Rank Math (WordPress), Yoast SEO (WordPress), Webflow SEO settings. Easiest path for non-technical owners
  2. AI coding tool. Claude Code generates JSON-LD from page content on demand. Paste into page head
  3. Manual paste from a schema generator. merkle.com, technicalseo.com, schema.org tools. Works but more tedious (developers.google.com)

5. Time: 30 hours over 6 weeks

Not all at once. The rhythm that works for a busy business owner:

  • Week 1: 4 hours on the question list and pillar page
  • Weeks 2 to 5: 5 hours per week writing 4 spokes per week
  • Week 6: 5 hours on schema, internal linking, and publishing

If you cannot commit 4 to 5 hours a week for 6 weeks, the launch will stall. Better to delay until a less-busy period than to start and abandon.

The 4 nice-to-haves

Adding these improves results but they are not blockers.

  1. Citation tracking tool. Otterly, Peec, Athena HQ at $50 to $150 per month. Skip for the first 60 days since there are no citations yet. Add at day 90 when baseline data becomes useful (otterly.ai)
  2. llms.txt file. 20 to 45 minutes to write, uncertain adoption in 2026, cheap to add. Bonus, not essential
  3. Bing Webmaster Tools account. Free. Helps ChatGPT Search find your pages faster. 15-minute setup
  4. A simple dev environment. If you use Claude Code, having a proper terminal and a local git repo saves time vs pasting files through a browser

The 4 things that look required but are not

Do not let these block the launch.

  1. A redesigned website. If your current site works, write the pages and launch. Design improvements come later
  2. A separate AEO subdomain. Some tutorials suggest aeo.yourdomain.com. Bad idea. Cluster at the root (yourdomain.com/[cluster-slug]/) builds more authority
  3. Massive backlink outreach. AEO cares less about backlinks than Google does. Skip the outreach grind for the first cluster. One or two real editorial mentions are plenty
  4. An analytics stack with dashboards. GA4 does not track AI citations anyway. Manual testing and a simple spreadsheet work fine for the first 90 days

The pre-launch checklist

Before you write the first page:

  1. Domain and hosting live, site renders to Google cache
  2. Question list of 20, locked, printed or saved somewhere visible
  3. AI writing tool subscription active
  4. One schema approach chosen (plugin OR AI coding tool OR manual)
  5. Calendar blocked for 4 to 5 hours per week for 6 weeks
  6. A spreadsheet or Notion doc to track draft status, publication date, and citation results per page

Checklist takes 2 hours to complete before you write word one. Cheaper than starting without it and discovering a blocker at week 3.

Common launch blockers

Four blockers that stall 80% of DIY AEO projects.

  1. The question list is never finalized. The owner keeps adjusting it. Pick 20 and commit. Revise after cluster 1 is done, not during writing
  2. Waiting for the "perfect" AI writing tool. The tools are fine. Use what exists. Switching mid-project costs weeks
  3. Schema anxiety. Spending 3 weeks researching schema options instead of writing. Pick the plugin or use Claude Code, move on
  4. Perfectionism on individual pages. Spending 6 hours on one page instead of 1.5 hours. The cluster effect matters more than any single page. Move faster, refine later