A history of SEO
From keyword-stuffing to semantic search
- Early days (1990s): The “Wild West” of the internet was driven by pioneer search engines like Yahoo! and AltaVista. SEO comprised submitting your site’s URL to web directories, basic keyword optimisation, meta tags, and on-page keyword stuffing.
- The rise of Google (early 2000s): Google’s PageRank algorithm changed directory search by treating the sites that linked to your site as a “vote” of authority. By ranking hyperlinks, Google made backlinks crucial to site ranking.
- Content and quality (2010s): In response to content farms and spammy link building that manipulated PageRank, Google shifted focus toward prioritising user experience and high-quality content. Social media engagement on platforms such as Twitter and Facebook was incorporated as a ranking indicator along with local SEO from Google Places (now Google Business Profile).
- LLMs and AI (2020s): In response to the rapid rise and evolution of AI-driven tools, modern search engine optimisation focuses on user intent, voice search, mobile-first indexing, and Generative Engine Optimisation (GEO).
Where we are now: Best practice SEO in 2026
Search engine optimisation in the age of LLMs
Large Language Models (LLMs) are a type of so-called Artificial Intelligence (AI) trained on datasets to understand and generate human-like text.
LLMs use Natural Language Processing (NLP) to match these datasets to preset patterns and so provide an appropriate response.
Models such as OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, Meta’s LLaMA, and DeepSeek power chatbots, code generation, content creation, and analysis.
LLMS have turned search into chat
Most recently, LLMs have transformed search from process of keyword-matching to a conversational, answer-driven process.
While traditional search retrieves links, LLMs aggregate information from multiple sources to provide a single answer.
This search process is called “semantic search”: information retrieval that understands the contextual meaning, intent, and relationship behind a query rather than just matching keywords.
This ability to “understand” language further allows LLMs to interact with users by allowing them to ask follow-up questions.
AI search has reduced human click-through rates
Because AI provides users with summaries to search queries, fewer users are clicking through to websites, reducing visitor numbers to those sites’.
The goal of SEO has therefore evolved to being cited AI summaries.
The new visibility tactics to increase visibility in AI-generated answers have been labelled Generative Engine Optimisation (GEO).
GEO is the new SEO
Generative Engine Optimisation (GEO) aims to maximise a site’s visibility in AI-driven search engines.
It’s the process of optimising content to be cited, referenced, or used as a foundational source for AI-generated answers.
GEO vs. traditional SEO:
- SEO targets Google’s top-10 blue links; GEO targets inclusion in direct, summarised answers.
- SEO emphasises keyword density and backlinks; GEO emphasises topical authority, clarity, and citations.
- SEO directs users to a list of websites; GEO provides a single summarised response.
How do I write for GEO?
Keep it chatty, and ask questions
Because AI search focuses on providing solutions to problems, content should anticipate and address user queries.
- Analyse AI Responses: Study what sources current AI engines cite for your target topics to identify gaps.
- Use conversational language. More users are searching by speaking, and listening to answers as audio. Write how you think your audience might phrase their questions.
- Structure copy as direct answers, lists of items, and actionable advice.
- Focus on long-tail keywords that reflect intent-based searches, e.g. “How do I do X?” “What are the differences between X and Y?” and so on
How do I structure content for AI search?
Establish a hierarchy of information
Content for LLM-based search is all about building and connecting clusters of information.
- Establish authority by producing comprehensive content that covers entire topics, with sub-topics branching off that structure.
- Create “pillar pages” for broad terms related to areas you want to prioritise, and link those pillar pages to smaller, more specific articles.
- Internal links between the pages further establish the hierarchy of information. Always use descriptive anchor text (e.g. “read our guide on X”) rather than “read more”.
Webstruxure works with clients to develop a logical “hub and spoke” heirarchy for their web content with clear navigation and internal links. We optimise header tags and implement technical solutons including schema markup to optimise your site content for AI search.
How important is the site build for GEO?
Fast, well-structured sites are better for AI
Technical optimisation is key to getting better results for generative text search engines.
AI systems is improved on by the following:
- How fast meaningful content appears.
- How easy it is to extract content.
- How stable the layout is.
At Webstruxure, a key focus of our website development, design and hosting solutions is optimising site performance.
For instance, we might recommend that sites load visible content quickly, have clean HTML (simple page structure, low code complexity), and are not JavaScript-heavy or blocked (i.e. have restricted access).
If a site cannot accommodate such restrictions, we’ll work with you to improve its SEO and GEO visibility in other ways.
