Pre-Launch Checklist: SEO, AIO, GEO, AEO, SXO & LLMO

I used to launch websites with a basic SEO checklist: sitemap, title tags, canonical URLs, done. Then I'd wonder why pages that ranked fine weren't showing up in AI Overviews. Or why Perplexity never cited me. Or why traffic from Google AI Mode looked completely different from classic search.

The problem wasn't the content. It was that I was optimizing for one surface while six others were deciding who gets cited, quoted, and found.

The core ranking factors haven't changed. E-E-A-T (Experience, Expertise, Authority, Trust), technical health, high-quality content that matches search intent, and authoritative backlinks with strong internal linking, these still drive rankings in 2026. This is the same MarTech stack discipline as everything else, what's changed is the number of surfaces you need those signals to work on.

Search in 2026 isn't one thing. It's SEO, AIO, GEO, AEO, SXO, and LLMO, six overlapping strategies that all want something slightly different from your page. Here's the pre-launch checklist I now run through before anything goes live.

What these six acronyms actually mean

Before the checklist, a quick map so none of this blurs together:

They overlap constantly. Good content that satisfies all six is just good content, clear, structured, fast, original, and trustworthy. But each strategy has a few non-obvious requirements. That's what this checklist is for.

SEO: The technical foundation still matters

Nothing else works if crawlers can't find your pages. This is the unsexy prerequisite, get it right before anything else.

Technical health

• robots.txt: Allow crawlers on important pages. Block staging, admin areas, search result pages with ?q= parameters.
• XML sitemap: Generated, only includes 200-status canonical URLs, submitted to Google Search Console and Bing Webmaster Tools on day one.
• HTTPS sitewide: SSL installed, all internal links and resources served over https://, HTTP redirects to HTTPS.
• Clean URLs: Lowercase, descriptive, hyphens not underscores. /seo-checklist not /page?id=123.
• Canonical tags: Self-referencing canonicals on every page. Prevents duplicate content issues you didn't know you had.
• 404 page: Custom page with navigation and a search bar. Don't make a lost visitor work to find you again.
• Redirects: If migrating from another domain or changing URL structure, 301 redirect map ready before launch.

Core Web Vitals

Test your scores with PageSpeed Insights, it shows both lab data and real-world field data from the Chrome User Experience Report.

• LCP under 2.5s: Largest Contentful Paint. Optimize hero images, use a CDN, defer non-critical resources.
• INP under 200ms: Interaction to Next Paint. Minimize heavy JavaScript on load.
• CLS under 0.1: Cumulative Layout Shift. Set width and height on every image. Reserve space for ads before they load.

On-page SEO

• Title tags: Unique per page, primary keyword included, under 60 characters. Google truncates anything longer on mobile.
• Meta descriptions: Unique, under 155 characters, written like ad copy, what's in it for the reader.
• Single H1: One per page, contains the primary keyword naturally.
• Logical heading hierarchy: H1 → H2 → H3. Don't skip levels for visual styling; do it in CSS instead.
• Image alt text: Descriptive and keyword-relevant, not stuffed. "blue running shoes on white background" beats "shoe1234".
• Internal linking: Every page reachable within three clicks from homepage. No orphan pages.

AIO: Optimizing for AI Overviews and AI Mode

Google runs two separate AI surfaces now, and they work differently. Understanding the difference matters before you optimize.

Google AI Overviews vs Google AI Mode

AI Overviews are the AI-generated summaries that appear at the top of standard Google search results. They pull from indexed web content and cite sources with small links. You still see traditional results below them.

AI Mode is a full conversational search experience, a separate tab in Google Search that works more like ChatGPT. It generates longer, multi-turn answers and draws on a wider range of sources. Think of AI Overviews as a summary layer on top of classic search, and AI Mode as a replacement for it.

The good news: the content signals that get you cited in AI Overviews generally work for AI Mode too. Clear structure, direct answers, and factual density transfer across both.

What AIO actually needs

• Structured data (JSON-LD): Organization, Article, FAQ, and Product schemas. AI parsers use schema to understand entities without reading every word.
• Clear definitions: Start sections with one-sentence definitions. "AI Overviews are Google's AI-generated summaries that appear above traditional search results." AI loves extracting clean definitions.
• Answer-first structure: Give the direct answer in the first two sentences under each heading. Everything after that is context.
• Author authority: Detailed author bio, LinkedIn link, credentials visible. AI uses authorship to gauge trustworthiness.
• About page: Complete with contact info, mission, editorial standards. AI Overviews penalize sites that look anonymous.
• E-E-A-T signals: Experience, Expertise, Authoritativeness, Trust. Show who wrote it, when, why they know this.

GEO: Getting cited by ChatGPT, Perplexity, and Claude

GEO is about standalone AI engines, tools people use instead of Google, not alongside it. Perplexity, ChatGPT Search, Claude.ai. These systems crawl the web independently and choose sources based on different signals than Google does.

The core concept here is information gain, does your content add something new? LLMs deprioritize pages that just restate what ten other pages already say. They reward original data, unique insights, and expert perspectives that can't be found elsewhere.

GEO checklist

• Original data or research: Your own numbers, survey results, or case study findings. Even a small dataset you collected is worth more than citing someone else's.
• Named expert quotes: Direct quotes from people with credentials and names. LLMs cite quotes because they're specific and attributable.
• Contrarian perspective: Don't just summarize the consensus. If you disagree with common advice, say so with a reason. Unique perspective gets cited; recaps don't.
• Outbound links to authority sources: Link to .gov, .edu, major publications. LLMs use your link neighborhood to gauge research quality.
• Specific numbers and dates: "Conversion rates dropped 23% in Q1 2026" beats "conversion rates dropped recently." Factual density is a GEO ranking signal.
• Brand mentions off-site: LLMs cross-reference your brand against their training data. Be mentioned on authoritative sites before you launch.

AEO: Winning featured snippets, PAA, and voice search

Answer Engine Optimization is about owning the direct-answer real estate, Featured Snippets, People Also Ask boxes, and the results that voice assistants read aloud. These surfaces reward a specific writing pattern.

The inverted pyramid format

Answer the question in the first 40–60 words. Then elaborate. Then add context. Voice assistants and snippet algorithms read the opening block first, if your answer is buried in paragraph three, you don't get cited.

AEO checklist

• Question-shaped headings: "How does X work?" "What is the difference between X and Y?" Match the literal search query in your H2.
• 50-word answer blocks: Direct answer right below each question heading. This is your snippet bait.
• FAQ sections: Four to five specific questions per page with direct, declarative answers. Not "it depends" answers, actual answers.
• FAQ schema: Apply FAQPage JSON-LD to all FAQ sections. Increases rich result eligibility.
• Tables for comparisons: Snippet algorithms and voice assistants love structured data. If you're comparing things, use a table.
• Numbered lists for processes: How-to content performs better in numbered steps than prose paragraphs.
• PAA research: Pull the People Also Ask questions for your target keywords. Answer those exact questions explicitly, don't paraphrase them.

SXO: The experience after the click

Search Experience Optimization is the bridge between ranking and conversion. It asks: once someone finds you, do they stay? Do they do the thing you want them to do?

Google uses engagement signals (dwell time, bounce rate, scroll depth) as soft ranking inputs. More importantly, traffic that doesn't convert is just infrastructure cost. SXO is where rankings become revenue.

SXO checklist

• Above-the-fold test: H1, core value prop, and primary CTA visible on mobile without scrolling. If a visitor has to scroll to understand what you do, you've lost them.
• Readability: Short paragraphs (two to three sentences max). High color contrast. Font size at least 16px. No walls of text.
• Navigation clarity: Menu labels should be obvious, not clever. "Work With Me" beats "Collaboration Opportunities."
• Clear CTA per page: One primary goal per page. Buy, subscribe, book, download. Don't make visitors guess what to do next.
• No dead ends: Every page should lead somewhere. Related content, next step, contact, no page should be a dead end.
• GA4 installed: Key events tracked and verified before launch, form submissions, button clicks, outbound link clicks.
• Microsoft Clarity or Hotjar: Session recording and heatmaps tell you what analytics can't. Install on day one; watch the first 50 sessions.
• Accessibility basics: Keyboard navigation works. Images have alt text. Color contrast passes WCAG 2.1 AA. Screen reader tested.
• Mobile tested on a real device: Not just Chrome DevTools. Touch targets, form inputs, tap behavior, these only show up on real hardware.

LLMO: Making your content easy for AI to ingest

Large Language Model Optimization is the newest strategy on this list, and the one most people ignore completely. LLMO is about content engineering that's friction-free for AI bots to crawl, read, and store in their knowledge base.

This matters because LLMs don't just search, they learn. If your content is hard to parse, AI tools train on worse data about your topic and represent you poorly (or not at all) in generated answers months later.

LLMO checklist

• Whitelist AI crawlers in robots.txt: Explicitly allow the bots that power AI search:
  • OAI-SearchBot, ChatGPT Search
  • PerplexityBot, Perplexity
  • ClaudeBot, Anthropic
  • Google-Extended, Gemini / Google AI
  • Applebot-Extended, Apple Intelligence
• Create an llms.txt file: Drop a plaintext file at yourdomain.com/llms.txt, a Markdown-formatted index of your most important pages, written for AI consumption. It's like a robots.txt but for LLMs. Emerging standard but already adopted by Perplexity and others.
• Server-side rendering: Core content must be in the HTML source, not rendered by JavaScript. Lazy AI scrapers skip JS execution. If your H1 and first paragraph are in a React component that hydrates client-side, many bots will miss them entirely.
• Answer-first paragraphs: Lead with the direct answer before context and background. LLMs process text in segments, if the answer is buried, the model's window may close before reaching it.
• Reduce pronoun ambiguity: Don't say "it" or "they" when referring to your brand or product in consecutive sentences. Repeat the entity name. LLMs lose track of subject when parsing isolated text blocks.
• Semantic HTML: Use <table>, <ul>, <ol>, <dfn> for definitions. LLMs parse semantic structure faster than prose. A <table> tag signals "here is structured comparison data" without the model having to infer it.
• Hard freshness timestamps: Visible datePublished and dateModified in both the page text and schema markup. LLMs scoring for time-sensitive queries filter out content with no explicit date.
• Consistent off-site presence: Your brand description, name spelling, and positioning should match across your site, LinkedIn, Crunchbase, industry directories. LLMs build a knowledge graph by cross-referencing, inconsistency creates gaps.

Code examples: the three files most sites get wrong

Most sites skip these or copy a generic template and forget to update it. Here's what each file should actually look like, including the AI crawler entries that most guides don't mention.

robots.txt, whitelist AI crawlers explicitly

Default robots.txt tutorials block everything or leave out AI bots entirely. The version below allows all major crawlers, including the AI agents that power ChatGPT Search, Perplexity, and Google's Gemini, while blocking admin and staging paths.

# Allow all major crawlers including AI bots
User-agent: *
Allow: /
Disallow: /admin/
Disallow: /staging/
Disallow: /*?* # block ?q= style search result pages

# AI search crawlers, allow explicitly
User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: Applebot-Extended
Allow: /

Sitemap: https://yourdomain.com/sitemap.xml

llms.txt, an index for AI crawlers

Place this file at yourdomain.com/llms.txt. It's a plain Markdown document that tells AI crawlers which pages matter most and what each one is about. Think of it as a table of contents written for machines. Perplexity and a growing number of AI search tools already read it.

# Your Site Name

> One-sentence description of what your site covers and who it's for.

## Core pages

- [About](https://yourdomain.com/about): Who runs this site and their credentials.
- [Services](https://yourdomain.com/services): What you offer and to whom.
- [Contact](https://yourdomain.com/contact): How to get in touch.

## Key content

- [Post title here](https://yourdomain.com/blog/post-slug): One-line summary of what this page covers.
- [Another post](https://yourdomain.com/blog/another-slug): One-line summary.

## Optional: exclude noisy pages

- /tag/: Tag archive pages, low value for AI crawlers.
- /search/: Internal search results, not useful to ingest.

Structured data (JSON-LD), schema for Article + FAQ

Add this inside a <script type="application/ld+json"> tag in your page's <head>. The Article schema tells search engines and AI crawlers who wrote it and when. The FAQPage schema makes your FAQ section eligible for rich result display in Google.

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Article",
      "headline": "Your Page Title Here",
      "description": "Your meta description here.",
      "datePublished": "2026-06-25",
      "dateModified": "2026-06-25",
      "author": {
        "@type": "Person",
        "name": "Your Name",
        "url": "https://yourdomain.com/about"
      },
      "publisher": {
        "@type": "Organization",
        "name": "Your Site Name",
        "url": "https://yourdomain.com"
      }
    },
    {
      "@type": "FAQPage",
      "mainEntity": [
        {
          "@type": "Question",
          "name": "Your FAQ question here?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Your direct answer here. Keep it under 300 characters for snippet eligibility."
          }
        }
      ]
    }
  ]
}

Launch-day validation: the final sweep

Before flipping the switch, run through this list in order. It takes about an hour and catches the things you'll spend three weeks debugging post-launch.

1. Crawl with Screaming Frog or Sitebulb. Look for 404s, redirect chains, orphan pages, missing title tags, and duplicate H1s. Fix before launch, not after.
2. Run the Rich Results Test. Paste each schema-rich page URL into Google's Rich Results Test. Errors here mean your structured data won't render as rich snippets.
3. Lighthouse audit. Target 90+ on Performance, Accessibility, Best Practices, and SEO. Run it in Chrome DevTools or use PageSpeed Insights for a real-world score that reflects actual user data, not just your local machine.
4. Mobile-friendly test. Google Search Console has one. Run it. Don't assume DevTools gives you the real picture.
5. Verify Analytics and Search Console. Confirm GA4 is firing, conversion events are tracked, and Search Console property is verified. Do this the day before launch, not after.
6. Submit sitemap to Google Search Console and Bing Webmaster Tools. Don't wait for Google to discover your pages organically. Submit on launch day. Then ping IndexNow to instantly notify Bing and other supported engines.
7. Test every form and CTA. Submit test entries. Check the confirmation emails. Confirm the thank-you redirects work. Form failures are silent revenue killers.
8. Check staging blocks are removed. Your staging environment should have noindex and password protection. Make sure those aren't on production.

The 80/20 version if you're launching fast

If you're shipping in 48 hours and need to prioritize, do these ten things and worry about the rest in week two:

1. Technical SEO health (sitemap, robots.txt, HTTPS, canonical tags)
2. Page speed passing Core Web Vitals
3. Clean site architecture with internal linking
4. Quality content that directly answers search intent
5. Organization + Article schema on key pages
6. FAQ sections with FAQPage schema
7. Author bio with credentials and LinkedIn link
8. GA4 + Search Console + Bing Webmaster Tools configured
9. Answer-first paragraph structure throughout
10. At least one piece of original data or insight per key page

What I'll keep updating

This checklist will change. Google AI Mode is still rolling out. Bing's AI Performance data is new. The Generative AI report in Search Console only landed recently. LLMO as a formal strategy barely has a name yet.

What won't change: the underlying principle. AI systems and search engines both reward content that is clear, structured, honest, fast, and original. That has been true for fifteen years and it's more true now than ever.

Build for the human first. The machines follow.

Frequently asked questions