Retrieval-Augmented Generation (RAG)

Why RAG Exists

LLMs are trained on data up to a certain point in time, which means they can become outdated. RAG solves this by allowing the model to search for and retrieve current information before generating a response. When you ask Perplexity or ChatGPT (with browsing enabled) a question, the model first searches the web, retrieves relevant pages, and then generates a response that incorporates this fresh information. This is RAG in action.

How RAG Impacts Brand Visibility

RAG is extremely important for brand visibility because it creates a direct path from your website to AI responses. When an AI model uses RAG, it actively searches for and retrieves content from the web. If your content is well-structured, authoritative, and relevant, it is more likely to be retrieved and incorporated into the AI's response — complete with a source citation linking back to your site.

RAG-Powered AI Products

Several major AI products use RAG: Perplexity is built entirely around RAG, always searching the web before answering. Google AI Overviews use RAG to incorporate current search results. ChatGPT with browsing uses RAG when it needs current information. Microsoft Copilot combines Bing search with RAG to provide up-to-date responses. Each of these products presents opportunities for your brand to be cited as a source.

Optimizing for RAG

To be picked up by RAG systems, your content needs to be easily discoverable and clearly structured. Use descriptive headings, include relevant keywords naturally, provide factual data and statistics, and ensure your pages load fast and are crawlable. Implementing llms.txt and structured data can also help AI crawlers understand and retrieve your content more effectively.