What is AI Search?

Generative AI search is a turning point in how users engage with search engines. Unlike conventional search methods that return a list of links, AI search synthesizes answers directly by leveraging large language models (LLMs). Instead of sifting through multiple sites, users receive cohesive, conversational responses.

Take Google’s AI Overviews, introduced in 2024, as an example. For queries like “best Italian restaurants in New York,” users now see a summary highlighting top picks and unique features without needing to click multiple links. Similarly, Arc Browser integrates AI-driven page summarization, helping users quickly extract key points from lengthy content, and an agentic AI architecture to search vast amounts of sources to return comprehensive answers rather than a list of web pages.

For businesses, this change poses a pressing question: How do you stay visible when users aren’t browsing traditional search results?

The New SEO Landscape: 2 Emerging Approaches

To adapt, companies are exploring two core strategies:

1. Generative Engine Optimization (GEO) GEO represents the most actionable and immediate strategy. By structuring content to align with AI search models, companies can ensure their information surfaces in AI-generated responses. To simplify, a new approach to SEO takes into account how LLMs generate and prioritize content.

Practical GEO Strategies

• Content Structuring Break content into clear, digestible sections with strong headings. AI models favor organized, well-annotated content.
• Keyword Optimization Traditional keyword research still applies but must now reflect user intent. Phrasing content as direct answers to questions increases visibility.
• Structured Data: Utilizing schema markup and structured data formats helps AI systems better interpret the content and present information in a user-friendly way.
• Credibility and Source Markers Link to reputable sources and provide verifiable claims. AI systems prioritize authoritative and accurate content.
• Engagement and Updates AI models favor frequently updated content. Refresh older posts with new insights to maintain relevance.

2. Providing Data for Fine-Tuning Enterprises can collaborate with LLM providers (like OpenAI or Anthropic) to fine-tune models on proprietary datasets. While effective, fine-tuning requires significant resources and high-quality data. Companies would need to take a careful approach to identifying what non-confidential data should be provided and invest in establishing a partnership with likely multiple model providers. While these partnerships exist, for example, companies providing data on niche languages to ensure that LLMs can generate and translate these rare idioms, a commercial model does not yet exist for enterprises that want to use this as a marketing strategy.

The Future of SEO: Pay-to-Play?

Fine-tuning and influencing LLMs hint at a future where enterprises pay to ensure their content dominates AI-generated responses. Could this lead to a model where visibility is restricted to those who can afford it? Or will open-source initiatives balance the scales?

As generative AI search grows, companies must proactively adapt. How will your SEO strategy shift to align with the AI-driven search landscape? Are you already experimenting with GEO techniques or fine-tuning models?

T. Babucci