Introduction
Generative engine optimization has a content side and a technical side. The technical side decides whether AI crawlers can reach your pages, whether engines can tell what each page is about, and whether the facts on it are easy to extract. This playbook lays out a technical GEO rollout in a sensible order: crawler access, a baseline site audit, llms.txt, schema (JSON-LD), FAQ markup and verification with real crawler data.
Who it is for: technical SEOs, web developers and marketing engineers who own templates, robots.txt and structured data, and want one plan the content team can rely on.
None of these steps guarantees a citation. Together they remove the technical reasons an engine might skip or misread a page, so your content gets a fair chance.
The challenge
Technical GEO is a collection of small files and tags, often owned by different people. robots.txt sits with one team, the CDN with another, schema comes from an SEO plugin, and nobody owns llms.txt. Changes are made one at a time, without a baseline or a check afterward.
Mistakes also fail quietly. A staging robots.txt with a blanket disallow goes live, a firewall rule blocks a search bot, two plugins output duplicate schema, or a page cache serves a stale file. None of these shows up as an error in your analytics, because bots do not run your analytics script.
The approach
Roll out in layers, from access to meaning, and measure each layer. Start with a handful of representative URLs per template, fix at the template level, then re-audit.
- Site Audit checks a page against 49 signals in five categories (Structure, Content, Authority, Trust and E-E-A-T), scores it out of 100 and writes fixes. Audits per month: 50 on Shine, 300 on Signal and unlimited on Scale.
- The GEO tools on Signal and Scale are an llms.txt optimizer, a Schema (JSON-LD) generator and an FAQ generator with FAQPage markup.
- Free tools cover the basics on any plan: a robots.txt tester, a robots.txt AI generator, an llms.txt generator, schema generators and a schema markup validator.
- AI bots analytics, on Signal and Scale, shows which verified AI bots read which pages, from your server logs.
- On WordPress, the free plugin serves llms.txt and llms-full.txt, Markdown versions of pages, schema without duplicates and FAQ blocks.
Step by step
- Inventory your templates: list the page types that matter, such as home, product, pricing, docs, blog posts and comparison pages. Pick two or three representative URLs for each.
- Check crawler access: paste your live robots.txt into the free robots.txt tester and test each representative URL against the AI crawlers you want, such as search and user-fetch bots. Then confirm your CDN or firewall does not block them. Use the robots.txt AI generator if you need to rewrite the rules.
- Run a baseline audit: audit each representative URL in Site Audit. An audit usually takes about 30 seconds. Note the failed signals per template, since a failure on one product page usually means a failure on all of them.
- Fix structure at the template level: work through the high-priority AI Recommendations first: one clear H1, descriptive headings, a direct answer near the top and a useful meta description. Recommendations include ready-to-paste drafts. Replace any Your input needed placeholder before you ship.
- Publish llms.txt: create a short llms.txt file at the root of your domain that names your site, summarizes what you do and links to your most important pages with a one-line description each. Use the llms.txt optimizer on Signal and Scale, or the free llms.txt generator.
- Add sitewide Organization schema: describe your company once, with name, logo, URL and official profiles, and keep it identical across templates.
- Add page-level schema: generate JSON-LD for each template, such as Article for posts, with the Schema generator on Signal and Scale or the free generators. If an SEO plugin already outputs schema, extend it rather than adding a second copy.
- Add FAQ markup where it fits: on pages with a visible FAQ, add FAQPage markup that matches the visible questions word for word, with the FAQ generator or the free FAQ schema generator.
- Validate everything: paste each template’s JSON-LD into the free schema markup validator and fix errors before rollout.
- Turn audit failures into actions: on any page that still fails checks, click Send to Action Center to create one Fix action with tasks. Sending it again while the action is open will not create a duplicate.
- Verify with crawler data: on Signal or Scale, connect a log source to AI bots analytics and confirm that search bots read your key pages. The search-bot alert lists pages AI answers cited in the last 30 days that no AI search bot read in that time.
- Re-audit and record: re-run the audit on each representative URL, compare the score with your baseline, and document every rule and file so the next change does not undo it.
A minimal llms.txt file looks like this. Replace the example names and links with your own:
# Example Co
> Example Co makes scheduling software for small clinics.
## Product
- [Pricing](https://www.example.com/pricing/): plans and what each includes
- [Integrations](https://www.example.com/integrations/): calendars and tools we connect to
## Help
- [Getting started](https://www.example.com/docs/start/): first steps for new accounts
Tip: Only mark up what a reader can see on the page. Markup that does not match the visible content can be ignored, or counted against you.
Who owns what after rollout
Write down an owner for each layer: robots.txt and CDN rules, llms.txt, sitewide schema, page-level schema and FAQ markup. Add a short check to your release process, such as testing robots.txt after every deploy, so a template change or a new plugin does not quietly undo the work.
What to measure
- Site Audit score per template, and the site score over time.
- Failed signals by category, before and after the rollout.
- Validator errors per template, which should reach zero.
- AI crawler hits and last seen dates on your key pages, by bot type.
- Pages in the search-bot alert, which should shrink as access issues are fixed.
- Citations to the pages you updated, and result checks at 7, 14 and 30 days on Fix actions.
Technical fixes usually show first in the audit and crawler data, and later, if at all, in citations. Keep the content work going in parallel, because markup helps engines read a page but does not replace a clear answer.
Next steps
Read the guides on site audit, GEO tools and AI crawler analytics, and work through the AI search optimization checklist. Start with the free robots.txt tester, llms.txt generator and schema markup validator. The GEO tools and AI bots analytics need Signal or Scale, so check pricing, or start a 2-day free trial.


