You can't buy your way into most AI answers (yet), but you can influence whether and how you appear. The discipline is often called generative engine optimization (GEO) or answer engine optimization.
What the research says
A study by researchers from Princeton and others — GEO: Generative Engine Optimization (accepted at KDD 2024) — tested content changes on a generative search engine and found that methods such as adding citations, quotations and statistics could boost a source's visibility in responses by up to 40%, with effects that differ by domain. It is early research on test systems, but it points in a clear direction: specific, sourced, quotable content gets used.
What brands can do
- Publish clear fact pages. Products, specifications, pricing logic, policies, comparisons and FAQs — in plain language, with dates and sources.
- Answer real buyer questions. Write for the questions people ask assistants ("best X for Y", "X vs Y", "is X safe for…").
- Be citable. Original data, named experts, specific numbers with sources.
- Be present where AI looks. Reviews, reputable comparison and industry sites, press coverage, partner pages, community forums.
- Keep it technically readable. Crawlable pages, structured data, no key facts hidden in images or PDFs only.
- Correct errors. When assistants get you wrong, fix the sources they draw on.
Brand strength shows up in models too
In 2022, Heinz and its agency Rethink asked the DALL-E 2 image model to draw "ketchup"; the images looked like Heinz bottles, and the brand built its "A.I. Ketchup" campaign around it. Strong, consistent brand assets become part of what AI systems "know" about a category.
What not to do
Mass-produced thin content, hidden text aimed at AI crawlers or fake reviews. They damage trust, may break platform rules or law, and tend to be filtered out.