MerseySearch - Welcome to the Generative Information Retrieval Era and What It Means For SEO

Transcript

1. Welcome to the Generative Information Retrieval Era -
2. Helllooooo
3. And... pomeranians and running
4. What even is Generative Information Retrieval?
5. It comes from the scientific field behind search - the field of 'Information Retrieval' (IR)
6. It's 'generating' answers or some other output rather than referring searchers to sources in search using AI / machine learning.
7. From this...
8. To this...
9. It's a total paradigm shift – the biggest change in search EVER
10. For the time being Google generative IR directly in SERPs is... AI OVERVIEWS AI MODE
11. AI Mode extends upon AI Overviews
12. Google Gemini plays a big part
13. But it's still... Minimum Viable Product
14. Generative Information Retrieval challenges... • Difficult to tie generated text to sources • Often cites the wrong sources Unfaithful • Hallucinates – Fabricates non-existent entities • Gets facts that do exist wrong Unfactual
15. Yikes
16. Concerns in information retrieval field 'Involution not evolution' (Ricardo Baeza-Yates) Google are "rushing ahead" (Bender & Shah) Searchers still want to learn and search for themselves (Bender & Shah) Generative IR (e.g. AI Overviews) takes away the agency (control) of searchers Generative IR doesn't meet Belkin's -16 information seeking strategies 'Rethinking Search' paper by Google team rebuked by many reviewers
17. Why the rush?
18. Rethinking search with AI foundations Delphic costs (Broder) •Reducing the 'search' cost burden on searchers •Particularly Gen Z and Millennials tempted to defect to ChatGPT et al Rethinking search •Take the knowledge to the searcher rather than make them search
19. It's here right now... MVP
20. So, what should we do about?
21. DON'T be scared out of your mind
22. Machine learning has been part of Google Search for years already Re-ranking and late stage re-ranking Crawl scheduling via importance prediction and spam swerving Feature engineering optimisation (understanding which factors should work together) Pairwise, pointwise and listwise ranking comparisons Learning to rank Index pruning Loads more
23. Firstly...Be Realistic, but ready
24. AI Search is still largely inaccurate
25. But... We must prepare
26. The landscape is changing fast
27. All Google Generative IR applications are emerging rapidly
28. AI Search / LLMs convert too  Salt agency studies show AI search applications / LLM traffic converts at a similar rate to organic search S Source - https://salt.agency/blog/do-users-really-show- higher-intent-when-they-click-through-from-an-llm-to-a- website/
29. Secondly...Consistency in SEO is more key than ever... Here's why
30. Two types of generative information retrieval Closed book Information comes only from within the model Open book Information comes from within the model and also can come from external sources to provide grouding
31. Web pages / content provide the grounding
32. Open book (AI overviews e.g.) uses retrieval augmented generation (RAG) from search index for 'grounding'
33. None of this works without quality unambiguous 'context'
34. You guessed it... Context is our 'in'
35. Creating great context is just good SEO Relatedness Schema Semantic headings Remove ambiguity Consistency in clustering Clear topic chunking in long form content
36. Content maintenance with high intent understanding likely took a step up in importance too
37. Continuous content auditing https://about.ads.microsoft.com/en/bl og/post/april-2025/mastering-ai- powered-search-next-level-strategies-for- marketers
38. Fabrice (Bing) blog post Summarised by ChatGPT  Large Language Models (LLMs) Enhance Search Experience  LLMs improve search relevance and user interaction by understanding context and predicting user needs, leading to more intuitive and efficient search experiences.  Integration of AI Technologies for Seamless Search  Combining LLMs with smaller, specialized models and real-time search capabilities ensures users receive accurate and up-to-date information.  Content Planning for AI Search  Marketers should conduct comprehensive content audits, maintain content freshness, target high-intent queries, and structure content with schema markup to align with AI search requirements.  Creating AI-Optimized Content  Content should be clear, structured, and use natural language. Incorporating expert insights and formatting with FAQs, lists, and schema markup enhances discoverability and engagement.  Continuous Optimization for AI Search  Regularly update content, adapt to AI trends using analytics, perform A/B testing, and personalize content based on user behavior to maximize visibility and engagement.  Utilizing IndexNow for Content Discovery  Implementing the IndexNow protocol allows for quicker notification to search engines about content updates, ensuring faster indexing and visibility.  Staying Ahead in AI-Powered Search  Adapting content strategies to align with the capabilities of LLMs and AI technologies is crucial for maintaining relevance and competitiveness in the evolving search landscape.
39. Know Thy User... Because LLMs will increasingly LLMs will build a 'user memory' Increased personalisation We're creating content for machines becoming search assistants to humans they know
40. Tech SEO... You guessed it. Still matters  If you don't get indexed in Bing you don't show up in ChatGPT  If you block Google-extended you can't show up in Gemini or AI Overviews  Very few LLM crawlers render Javascript (Google Gemini does)… Avoid client side Javascript even more so  Avoid distractions... e.g. lots of debate about whether to use llms.txt in addition to robots.txt. Probably not worth your time for the moment  Schema, structured data and ontologies / knowledge graphs hugely important
41. Finally... Get measuring and claiming ALL the LLM traffic as organic / SEO
42. Anything not paid search or social channels is 'OURS'
43. Thank you X - @dawnieando Bluesky - @dawnieando Linkedin – Ms Dawn Anderson Bertey.com