2024 was the year Generative AI shook up Search. In October alone, OpenAI launched SearchGPT and Perplexity hit over 100 million search queries per week.

For marketers, this shift means added complexity and the need to get familiar with advanced AI models. While the long-term effects on click rates and website traffic are still uncertain, knowing how AI Search Engines work is essential for tailoring your SEO strategy.

In this article, we break down the differences between traditional Google Search and Conversational Search engines and share tips on how to keep your content relevant.

How Google’s Search Engine works

Generative AI might have changed the SEO game, but AI in Search isn’t new. Google has used AI algorithms to surface content for years. Here’s a quick refresher:

• Indexing: Search engines use crawlers to discover pages across the web and index them to create a searchable database. The easier a website is to crawl, the more likely it is to get featured in search results.

• Keyword Analysis: The algorithms analyse keywords to determine how well a webpage matches a user's query.

• Natural language processing (NLP): Google’s AI systems RankBrain and BERT relate words to concepts to deliver better results. It uses NLP to understand the intent behind a search query and then links it to similar known searches to predict the most relevant results.

• Multitask Unified Model (MUM): In May, Google introduced a Generative AI model, which they claim is 1000x better than BERT. MUM can both understand and generate language and is Multimodal, meaning it can interpret information beyond text (such as images via Google Lens).

• AI Overviews: This year Google rolled out a new feature that uses large language models (LLMs) to summarise search results. While there are still some question marks around how these are pulled, websites with high rankings are known to be referenced more frequently.

In addition, Google looks at backlinks, page experience, and user behaviour (e.g. click-through rates and time spent on pages) to determine if the content is valuable.

How AI Search Engines work

AI search engines take user intent a step further to deliver more personalised results.

• Smaller, Curated Index: Conversational search engines like Perplexity focus on quality by indexing the "head of the distribution curve" – the most popular and high-quality content from reputable sources.

• Intent Analysis: AI Models like GPT-4 Omni and Claude 3 help generative search engines understand complex, conversational language and the context behind queries.

• Content Quality: By evaluating trustworthiness, popularity, and freshness, it decides what content to present.

• Query Matching: Similarly to Google, conversational search engines look at relevance, language, and location to match users with the best content. It also learns from past searches to deliver personalised answers.

• Contextual Memory: Follow-up questions allow engines like Perplexity to recall the context of previous queries, making responses more relevant.

• Multimodal Analysis: AI engines can process content formats beyond text, including videos and audio, to provide richer responses.

• Machine Learning: By continuously learning, it refines and improves their answers over time.

Key differences

So what does this mean for SEO? Here are a few key differences worth paying attention to as marketers.

Tips for content creators 

The shift towards Entity-based SEO and NLP-powered search means context matters more than keywords.

1. Produce high-quality, original content: Create content that demonstrates expertise, authority, and trustworthiness (E-A-T). Focus on highly relevant, user-centric content to improve your odds of being ranked. 

2. Opt for depth over volume: Cover topics comprehensively, providing in-depth answers. Use well-researched sources and link to T1 publications.

3. Use conversational language: Use a natural, conversational tone in your writing. Anticipate potential follow-up questions and structure your content using FAQ-style subheadings and clear, accessible language. 

4. Be creative with content formats Utilise content formats like images, videos, infographics, and audio to boost engagement. This keeps users on your page longer, positively impacting SEO.

5. Write well-structured content: Organise your content with clear headings, bullet points, and concise language. Well-structured content is easier for AI algorithms to interpret and pull information from for search overviews and feature snippets. Adding structured data markup (like schema types such as Article, FAQ, HowTo) helps search engines better understand your content.

Measuring success

The ‘black box’ problem of advanced AI systems and deep learning models makes it difficult to learn how these systems decide what content to feature. Here are a few tips on how to gauge your SEO efforts despite the lack of transparency: 

• Search Engine Testing: Perform regular searches for your target keywords and see if your content appears in AI-generated summaries. 

• Google Search Console: Check your performance for queries that generate featured snippets—this can indicate your content is being summarised or highlighted.

• Google Analytics: Measure direct traffic from Perplexity and other AI Search tools. 

• SEO Tools: Tools like Ahrefs distinguish between AI-generated overviews and organic search results for more insights.

Continue tracking organic traffic, keyword rankings, and bounce rates. We watch this space to help you stay one step ahead.