Key Highlights:
- Google integrates LLMs in searches by using RAG to provide accurate AI answers with sources.
- Structured data remains essential for enhancing AI's understanding and improving search visibility.
- AI search data is included in Search Console reports, however no separate AI-specific reporting is planned yet.
During this week's Google Search Central live conference in Madrid, Google leaders shared some new insights on AI in search with the future of SEO. The meeting featured talks from Google's Search Relations team that included John Mueller, Daniel Weisberg, Moshe Samet, and Eric Barbera.
The whole coverage was done by Aleyda Solis, who also attended the conference and noted the main points.
Google's LLM Integration Architecture Revealed
John Mueller explained how Google uses Large Language Models (LLMs), through a method called Retrieval Augmented Generation (RAG), and using reliable information to build AI-powered search answers. According to Mueller, the process works in four steps:
Step 1: A user enters a question.
Step 2: The search engine finds the relevant information.
Step 3: The information is used to "ground" the LLM.
Step 4: The LLM creates an answer with supporting links.
This whole system is created to keep answers accurate and linked to their sources which addresses concerns about AI-generated errors.
No Special Optimization Required for AI Features
Google officially made it clear to SEO professionals that there are no extra tweaks needed for AI features. Here are the key points that Google shared in the conference:
- AI tools are still new and will keep on changing.
- User behavior is still growing with AI search.
- AI data appears with traditional search data in Search Console.
- Much like with featured snippets, there is no separate breakdown.
Sticking to your current SEO best practices is enough for now, but Google encourages you to report any unusual issues.
Structured Data Remains Essential in an AI World
During the conference, Google advised that despite advances in AI, structured data is important. Google suggested that you should:
- Keep using supported structured data types.
- For the right schemas, check Google's documentation.
- Understand that structured data makes it easy to read and index your content.
Even if AI can work with unstructured data, using structured data will give you a clear advantage in search results.
Controlling AI-Driven Presentations of Content
Google explains several ways to control AI features for website owners who are worried about their content:
Use the robots nosnippet tag to opt out of AI overviews.
Add a meta tag like <meta name="robots" value="non snippet">
Wrap certain content in a <AAdiv data-non snippet></div>
Limit the amount of textAshown with <meta name="robots" value="max-snippet: 42">
These options will work just like the controls of traditional search snippets.
Reporting & Analytics for AI Search
According to information shared by Solis, Google's approach to reporting was also discussed:
- AI search data is included with overall Search Console data.
- There is no separate report just for AI features.
- Releasing AI data separately might cause more confusion for users.
- Currently, there are no plans to report Gemini usage separately due to privacy issues, though this might change if new patterns are seen.
LLMs.txt and Future Standards
A potential field called LLMs.txt was also discussed, which could work like robots.txt and control AI usage. Mueller said: "this file only makes sense if the system doesn't know about your website".
Since Google already has plenty of data about most websites, this extra layer might be unnecessary. For Gemini & Vertex AI training, Google now uses a user-agent token in robots.txt, which does not affect search rankings.
SEO's Continuing Relevance in an AI-Powered World
It was made clear in the conference that basic SEO work is still important. Here are key points that were discussed for SEO work:
- Core SEO tasks like crawling, indexing, and content optimization remain the same.
- AI tools added new capabilities to digital marketing rather than replacing old methods.
- SEO professionals have to use their skills in a changing landscape.
This message was to reassure that if you have strong SEO basics, you can adapt to new AI tools without completely changing your strategies.
Industry Implications
Solis's whole coverage depicts that Google focuses more on user needs while adding new features. The goal is to keep delivering quality content and solid technical foundations. Although AI brings new challenges, serving users well does not change.
Some challenges remain like not having separate reports for AI features. However, as these features grow, more precise data may soon be available.
For now, SEO professionals should continue using structured data by following their proven SEO practices, and keeping up with new trends & developments.
