Google’s Liz Reid discussed on the Bloomberg Odd Lots podcast how AI Mode and AI Overviews enable detailed, need-based query patterns that introduce new challenges for Google and significantly change search behavior, impacting SEO strategies. **Keyword Fragmentation in AI Search** Reid highlighted that users have long wanted to express longer, natural language queries but had to simplify their searches into short keyword phrases like “best restaurants in New York, ” even when their real needs were more specific—such as finding a vegan-friendly restaurant with availability for five people. For nearly 30 years, SEO has relied on keyword research: selecting target keywords and optimizing content accordingly. However, short keywords often carry hidden meanings. Google traditionally used user clicks not only to rank websites but also to interpret ambiguous queries, displaying results based on the most popular interpretation. Pages aligned with less popular senses of a keyword rarely ranked well. Now, with AI-based search, users express their information needs more fully via longer, natural language queries, which helps Google retrieve precisely what they want. Reid explains that AI Overviews allow users to state their real problems without translating them into computer-friendly keywords, empowering people and simplifying the search process.

For example, instead of typing “restaurants New York, ” a user might specify “a vegan-friendly restaurant in New York for five people, affordable and kid-friendly, ” which was difficult to communicate in past search formats. **Key implications include:** - Complex questions posed in AI Search might not be answerable by a single webpage. - Such complex queries are often unique and rare, potentially reducing the value of optimizing specifically for those phrases compared to other SEO tasks. - Websites are likely to compete for AI Overview space alongside others, emphasizing the importance of strong branding, relevant images, and videos to stand out. - Despite the prevalence of long queries, Google deconstructs them into smaller, precise keyword phrases, running those through classic search. It then selects top results from those to synthesize an answer, meaning traditional SEO focused on well-optimized, specific queries remains vital. **Addressing Real Needs** Reid noted that breaking down a complex natural language query into smaller components is a challenge for quality control in AI Search. Because users aren’t repeating similar keyword phrases as in traditional search, Google can’t cache results as efficiently, which affects latency and system performance. Still, this evolution empowers users by reducing the time spent searching, addressing their true needs more effectively. Reid said, “If you ask users when was the last time they spent 20 minutes searching instead of 2, most find search quite efficient. ” The goal of AI search improvements is to make lives easier by directly fulfilling users’ real needs. This approach challenges SEOs to audit web pages not just for keywords and technical issues but for how well the content fulfills specific user needs. When evaluating sites—such as one with indexing problems—it’s useful to ask, “What need does this page fill?” and “How is it distinct and better than other pages?” For further insight, watch Liz Reid’s full interview: *Google’s Liz Reid on Who Will Own Search in a World of AI. * *Featured Image by Shutterstock/TierneyMJ*