New Research Helps Your Website Get Quoted by AI Chatbots

For twenty years, getting found online meant ranking high on Google. That is ending. Tools like ChatGPT, Perplexity and Google's AI summaries now answer questions directly, so people often never click a link at all. For anyone who runs a website, writes articles or sells things online, the new goal is simple: get quoted inside the AI's answer. Researchers call this "Generative Engine Optimization" — basically, SEO (search engine optimization) rewritten for chatbots. The problem? These AI systems are black boxes. Nobody knows exactly why one page gets cited and another gets ignored.

A new paper from researchers Shilpa Ramakrishna and William Andreopoulos offers a clever workaround. Instead of requiring you to know what people are searching for, their method reads your existing document and guesses the questions it could answer. Then it spots the gaps — useful details your page is missing that a curious reader would probably want. Add those, and the AI becomes more likely to pull from your page. They call it QI-GEO, short for Query Implied Generative Engine Optimization.

The results are modest but real. Tested against two standard research datasets, the method improved objective scores by up to 15.9% and subjective scores by up to 17.6%. More importantly, it produced roughly twice as many new AI citations as it lost. In practical terms: your existing article becomes more quotable without you having to research keywords or rewrite everything from scratch.

There is one big catch. This is an early academic paper, not a finished product, and it was tested on research benchmarks rather than the messy real world. It also nudges content creators toward writing for machines instead of readers — the same arms race that made Google search results feel bloated and spammy. Still, for small businesses and independent writers worried about disappearing from AI answers, it is a useful sign that visibility is still something you can work on.