# How We Build SEO, AEO, and GEO for Real AI Visibility

> Our agency method for technical SEO, content SEO, structured data, SSR, llms.txt, AI visibility, off-page trust, backlinks, and the measurement behind it.

Source: https://www.truenorthmarketing.ae/en/blog/seo-aeo-geo-operating-system
Author: Vedant Achharya
Published: 2026-07-19
Updated: 2026-07-19
Category: Strategy & Insights
Tags: SEO, AEO, GEO, Structured Data, Technical SEO, Content SEO, AI Visibility, UAE, GCC
Publisher: True North Marketing (truenorthmarketing.ae)

## Article

SEO, AEO, and GEO are not separate tricks. They are three views of the same problem: can the internet understand, trust, and recommend your business accurately?

At True North, we build search and AI visibility as part of the website operating system. That includes technical SEO, content SEO, structured data, server-rendered content, sitemaps, `robots.txt`, `llms.txt`, markdown summaries, internal linking, page speed, off-page trust, backlinks, partner profiles, and measurement. It also includes restraint. If the content is generic, no schema type will make it authoritative.

The goal is not to look optimized. The goal is to make the business easy to understand, easy to verify, and easy to act on.

Search visibility starts with a useful page for a real decision. AI visibility starts with the same page made technically clear: server-visible content, explicit entities, structured data, internal links, and consistent off-page proof.

## How does people-first content become machine-readable?

Google's current guidance on [creating helpful, reliable, people-first content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content) is still the cleanest editorial test: content should be made for people, demonstrate experience and expertise, and leave the reader satisfied. That does not conflict with AEO or GEO. It is the foundation.

The mistake is writing for a crawler first and then trying to humanize it. That creates the obvious patterns: repeated intros, keyword-loaded headings, fake certainty, generic "ultimate guide" structure, and source lists added for decoration.

Our order is different:

1. define the reader decision;
2. write the useful answer from our direct experience;
3. add examples, limits, and implementation detail;
4. structure the page so search and AI systems can extract it cleanly;
5. connect it to the rest of the site through internal links and schema.

That is how promotional content can appear naturally. We do not interrupt a technical article to say "hire us." We explain how we solve the problem and link to the relevant [services](/en/services), [case studies](/en/case-studies), or [partnerships](/en/partnerships) only where it helps the reader verify the method.

## Why does rendering matter for technical SEO?

Critical content must exist in the initial HTML, before any JavaScript runs. If a crawler or an AI system has to execute your page to see the words, you have already lost reliability, because not every bot renders JavaScript, and the ones that do render it later, slower, and less completely. So rendering strategy is the first technical SEO decision, not an afterthought.

Three modes matter, and serious sites use all three deliberately:

1. **SSG, static generation.** The page is built to HTML at deploy time, then served from the edge. Fastest to load, cheapest to cache, and perfectly crawlable. This is the default for marketing pages, blogs, docs, and anything whose content does not change per request.
2. **SSR, server-side rendering.** The server builds the HTML on each request. The right call when content is dynamic or personalized but still has to arrive fully formed for crawlers and LLMs, not assembled in the browser.
3. **Hybrid and incremental, ISR.** Static where possible, server-rendered where needed, with individual pages regenerated on a schedule or on demand. This is how most real sites ship: static speed with dynamic freshness, and no full rebuild to update one page.

We build these on modern, engineering-grade frameworks. Next.js gives full control over SSG, SSR, ISR, route metadata, image optimization, and server functions. Nuxt gives the same rendering model for Vue teams. TanStack Start is excellent React tooling for app-heavy products that need full-document SSR and typed routing. Shopify Hydrogen, built on Remix, is the right layer for headless commerce when Shopify stays the source of truth. What they share is the point: the meaningful content, copy, headings, links, and structured data, is server-visible by default.

| Stack | SEO/AEO advantage | Watchout |
| --- | --- | --- |
| Next.js | Strong route metadata, SSG/SSR/ISR, server components, structured pages, custom APIs | Needs disciplined engineering ownership |
| Nuxt | SSR/SSG for Vue teams, content-friendly architecture, strong routing | Same need for clean content contracts |
| TanStack Start | Genuinely strong React tooling for app-heavy builds with SSR and typed routing | Overkill for a simple marketing site; reach for it when the product needs it |
| Shopify native theme | Merchant-friendly commerce, stable checkout, product and collection operations | Theme and app bloat can hurt speed |
| Headless Shopify or Hydrogen | Deeper content and frontend control around Shopify data | More cache, preview, API, and deployment ownership |
| Webflow | Fast visual publishing for simpler marketing sites | Complex app logic and structured SEO can become awkward |

This is why [stack selection](/en/blog/development-brainstorm-to-production) belongs inside SEO strategy. On a modern, server-rendered stack, good content is crawlable, fast, and easy to connect to conversion data. On a slow, plugin-heavy legacy build, the same content fights the rendering the whole way.

## Should structured data describe visible truth?

Structured data is not a place to hide sales claims. It should clarify the page's visible content and relationships.

Google's [structured data guidelines](https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data) are clear that structured data should represent content on the page. I treat JSON-LD as a trust surface. If the schema says something the user cannot verify on the page, it weakens the system.

For service and ecommerce sites, we usually inspect:

| Schema | Use when |
| --- | --- |
| `Organization` | The business identity, logo, sameAs profiles, contact, and official URL are clear |
| `WebSite` | The site has a canonical entity and search/discovery context |
| `WebPage` | Important pages need explicit name, URL, description, and relation to the site |
| `BreadcrumbList` | Page hierarchy should be machine-readable and match visible navigation |
| `Service` | Service pages describe real offers, scope, area, and provider |
| `Product` | Product details, availability, price, and merchant truth are present |
| `FAQPage` | Questions and answers are visible, useful, and not duplicated spam |
| `BlogPosting` or `Article` | Authored content has title, author, image, date, and topic clarity |
| `Review` | Reviews are real, visible, and follow the platform's rules |

For our own site, we also care that `llms.txt`, markdown mirrors, visible copy, and JSON-LD do not contradict each other. AEO/GEO is partly consistency engineering.

## How do you write AEO answers with real judgment?

Answer Engine Optimization works when sections answer real questions directly. It fails when every heading becomes a forced question and every paragraph sounds like a snippet farm.

The structure I like:

- direct answer first;
- when the answer changes;
- what we do in practice;
- what the reader should not assume;
- one internal link to the next useful page.

Example: "Should every page be indexable?" The answer is no. Index pages that deserve discovery. Noindex thin, duplicate, private, temporary, internal search, and low-value filter pages. Then connect that rule to the sitemap, canonical tags, CMS workflow, and measurement.

This is not anti-SEO. It is better SEO because it respects the reader's decision.

## What does GEO need beyond llms.txt and markdown?

Generative Engine Optimization is about making it easier for AI systems to identify the right source, parse the page, and summarize the business accurately.

The `llms.txt` convention, described at [llmstxt.org](https://llmstxt.org/), proposes a clean markdown file that helps language models find useful site context. I treat it as a supporting layer, not a magic file.

Our GEO stack usually includes:

| Surface | What it does |
| --- | --- |
| `sitemap.xml` | Lists canonical public pages for discovery |
| `robots.txt` | Communicates crawl permissions and sitemap location |
| `llms.txt` | Summarizes important pages for AI readers |
| Page-level `llms.txt` | Gives concise context for a route when useful |
| Markdown mirrors | Exposes clean article or page text without layout noise |
| JSON-LD | Clarifies entities, breadcrumbs, articles, services, products, and FAQs |
| Internal links | Shows topical relationships and next-step hierarchy |
| SameAs profiles | Helps connect the brand entity across trusted external surfaces |

The content has to match across those surfaces. If the article says one thing, schema says another, and `llms.txt` says a third, AI systems are more likely to summarize the brand badly.

## Is content SEO about architecture or blog volume?

Content SEO is not publishing more posts. It is creating the minimum set of pages that answer real buyer, operator, and search questions better than the alternatives.

I separate content into four jobs:

| Content type | Job |
| --- | --- |
| Service pages | Explain the offer, scope, proof, process, and next step |
| Case studies | Prove capability with named context, work performed, and honest limits |
| Blog posts | Teach a decision, objection, process, or technical judgment |
| Trust pages | Show team, partnerships, policies, contact, and business legitimacy |

This is where our agency context matters. A blog on voice agents should connect to [AI automation](/en/blog/ai-automation-digital-marketing), [development](/en/services#development), and [contact](/en/contact) only when the reader has enough context to act. A technical SEO post should connect to [case studies](/en/case-studies) when proof matters, not as a random CTA.

Google's [SEO starter guide](https://developers.google.com/search/docs/fundamentals/seo-starter-guide) still emphasizes useful content, clear links, descriptive titles, crawlability, and avoiding practices that make pages worse for users. That is not basic advice. It is the part many sites skip while chasing advanced tactics.

## How should internal links move the reader forward?

Internal links should not be sprinkled like seasoning. They should express the site's logic.

For a growth agency, that logic might be:

- a strategy article points to the [about page](/en/about) because operating model matters;
- a development article points to [web and software development](/en/services#development);
- a creative article points to the [studio](/en/studio);
- an SEO/AEO article points to [partnerships](/en/partnerships) when platform credibility matters;
- case-study mentions point to the [case studies index](/en/case-studies) or the exact public case.

The anchor text should describe the destination. "Read more" is weak. "How we choose website tech stacks" tells both the reader and the crawler what the link is doing.

## How does off-page SEO prove your entity is real?

Backlinks still matter, but the lazy version of link building is dangerous. Buying irrelevant links, submitting to junk directories, or producing guest posts nobody reads may create short-term noise and long-term risk.

I care about off-page signals that strengthen the entity:

| Off-page surface | Why it matters |
| --- | --- |
| Partner directories | Confirms real platform relationships where verified |
| Founder profiles | Connects expertise to real people |
| Client mentions | Shows the agency exists in real commercial contexts |
| Case-study references | Provides proof beyond a service page |
| Industry directories | Helps discovery when the directory has editorial value |
| PR and speaking | Builds authority around the people and the business |
| Useful tools or research | Earns links because the asset helps someone |

For TNM, our formal partner and certification ecosystem includes Google, Meta, HubSpot, Vercel, DigitalOcean, and Cloudflare. Tools we use, such as Shopify, WP Engine, PostHog, Typeform, Klaviyo, Pinecone, n8n, Make, Claude, and ElevenLabs, are part of the working stack, but I do not call them formal partnerships unless they are verified as such.

That distinction matters. Trust grows when claims are precise.

## Technical SEO includes speed, but not only speed

Core Web Vitals, image strategy, script loading, crawl health, redirects, canonicals, metadata, hreflang, schema, and sitemap hygiene all affect visibility. But technical SEO is not only a PageSpeed score.

For a serious audit, I check:

- indexable route inventory;
- canonical and alternate language rules;
- status codes and redirect chains;
- title and meta description uniqueness;
- structured data validity and visible-content match;
- image size, alt text, loading, and format;
- JavaScript required to read primary content;
- duplicate and thin pages;
- internal link depth;
- event and conversion tracking;
- page templates that carry paid or organic demand.

The [LaserYou case study](/en/case-studies/laseryou-nextjs-aeo) is a good example of the full shape: performance, treatment content, structured data, `llms.txt`, markdown mirrors, tracking, accessibility, and a dated PageSpeed proof point. The point is not that Next.js magically ranks. The point is that the rebuild treated content, semantics, performance, and measurement as one system.

## Backlinks and partnerships should support the same story

A good backlink strategy should make the same claim the website makes. If the site says the agency specializes in performance marketing, development, AI automation, and creative, external surfaces should support that reality.

That can include partner listings, trusted directories, client references, founder profiles, speaking pages, case studies, and useful educational content. It should not include irrelevant posts placed only for anchor text.

The anchor profile should look natural because the work is natural: brand name, founder name, service categories, case-study subjects, and useful resources. If every link uses a commercial keyword, it looks manufactured because it is.

## Measurement connects visibility to revenue

SEO can create traffic that never helps the business. AEO can produce snippets that never produce trust. GEO can create AI visibility that cannot be measured perfectly. That is why I connect visibility back to business records where possible.

I look at:

- indexed pages and crawl errors;
- query and landing-page movement;
- qualified organic enquiries;
- assisted conversions and revenue;
- lead quality from organic journeys;
- branded search growth after stronger proof;
- backlink quality and referring domains;
- AI system summaries when tested manually;
- whether pages reduce repetitive sales questions.

This connects SEO to [performance marketing](/en/services#marketing), [web development](/en/services#development), [AI automation](/en/services#ai-automation), and [creative](/en/studio). Search does not live in a corner. It touches the page, the proof, the offer, the CRM, and the follow-up.

## The SEO, AEO, and GEO checklist I trust

Before I call a site ready, I want this operating stack in place:

| Layer | Pass condition |
| --- | --- |
| Strategy | Each important page has a reader decision and business purpose |
| Rendering | Core content is SSR, SSG, or otherwise server-visible |
| Metadata | Titles, descriptions, canonicals, Open Graph, and hreflang are deliberate |
| Schema | JSON-LD matches visible content and uses explicit entity names |
| Content | First-person expertise, useful examples, limits, and no keyword stuffing |
| Internal links | Natural links connect services, blogs, case studies, partnerships, and contact |
| AI surfaces | `llms.txt`, route summaries, and markdown mirrors match the public truth |
| Technical SEO | Sitemaps, robots, redirects, index rules, speed, and crawl health are clean |
| Off-page | Backlinks, profiles, directories, partners, and proof strengthen the entity |
| Measurement | Organic and AI-assisted visibility connect to business outcomes where possible |

That is the difference between SEO as a checklist and visibility as an operating system.

If your site is on Next.js, Nuxt, Shopify, or a hybrid modern stack, the principle stays the same: make the truth visible, structured, fast, connected, and externally credible. That is how we build search and AI visibility without turning content into spam.

## FAQ

### What is the difference between SEO, AEO, and GEO?

SEO improves discovery and rankings in search. AEO makes answers clear enough for answer engines and AI Overviews. GEO prepares content and technical surfaces for generative engines that summarize, cite, or recommend sources. Strong sites treat them as one system: useful content, crawlable rendering, clear entities, links, and trust.

### Does AEO replace SEO?

No. AEO depends on SEO fundamentals. A page still needs crawlable HTML, good information architecture, internal links, helpful content, structured data, fast loading, and authority signals. llms.txt, FAQ blocks, and answer-style sections can help only when the underlying page deserves to be trusted.

### Why does SSR or SSG matter for AI visibility?

SSR and SSG expose core content, headings, links, metadata, and schema in the initial HTML. That helps search engines, AI crawlers, social previews, and accessibility tools understand the page without relying on client-side rendering. Client interactivity is fine, but the content that carries meaning should be server-visible.

### Which schemas should a business website use?

Use schemas that match visible content: Organization, WebSite, WebPage, BreadcrumbList, Service, Product, FAQPage, BlogPosting, Article, LocalBusiness where accurate, and Review only when reviews are real and displayed. Structured data should clarify truth, not hide extra claims inside JSON-LD.

### What is llms.txt and should every site have it?

llms.txt is a proposed convention for giving AI systems a concise map of important site content and machine-readable summaries. It is useful when maintained carefully, but it is not a ranking switch. It should mirror public pages, canonical URLs, policies, and markdown versions without contradicting the site.

### How do you write content that works for Google and AI systems?

Write for a real reader decision first. Then make the structure extractable: clear headings, direct answers, examples, limits, internal links, and visible authorship. Avoid keyword stuffing, fake expertise, and generic source blocks. Google repeatedly frames useful content around people-first value, not search-engine-first production.

### Do backlinks still matter?

Yes, but quality and relevance matter more than volume. Useful backlinks come from real mentions, partner ecosystems, directories with editorial value, client proof, PR, case studies, research, and genuinely helpful assets. Paid spam, irrelevant directories, and manipulative link schemes create risk instead of durable authority.

### How does off-page SEO fit into AEO and GEO?

AI systems and search engines both need external confidence signals. Off-page work helps when it confirms the entity: consistent brand profiles, partner pages, directory listings, client references, founder profiles, media mentions, and authoritative backlinks. The goal is entity confidence, not a backlink number.

### How do Next.js, Nuxt, and Shopify differ for SEO?

The SEO model is the same, but the rendering and control differ. Next.js and Nuxt give strong SSG, SSR, and ISR with route-level metadata and structured data, which is why we default to them for content and marketing sites. Shopify needs theme, product, collection, canonical, and JSON-LD discipline, and headless Shopify or Hydrogen when the storefront needs more frontend control. We build on modern, server-rendered stacks so the content is crawlable and fast by default, rather than fighting a slow, plugin-heavy legacy build.

### How do you measure whether SEO and AI visibility are working?

We measure indexed page quality, crawl health, rankings, qualified organic traffic, assisted conversions, branded demand, backlink quality, AI citation visibility where observable, and whether the site is summarized accurately by major AI systems. Traffic alone is not the result. Commercial action and entity trust are the result.