Monica Lam’s Stanford lab, OVAL, presented a new prompting method called STORM at a top AI conference. They found that asking a chatbot dozens of targeted questions—rather than a single broad one—yields articles about 25% better organized than those from standard AI prompts. Over 70,000 users are already using STORM to produce fully cited, Wikipedia-grade write-ups on topics the model has never seen before. STORM breaks down into five steps. First, gather 6–8 expert perspectives on your subject to cover all major angles. Second, include quoted or linked interviews from actual specialists. Third, draft a rigorous outline that forces the AI to stick to a clear structure. Fourth, write each section with evidence and citations in place. Finally, run a blind-spot red-team review to catch gaps or unchecked assumptions before finalizing the text. The team built a suite of five prompts you can drop into Claude (an Anthropic model) to automate the process. Each prompt handles one stage—perspective gathering, citation integration, outline creation, section drafting and red-teaming. According to OVAL’s tests, articles produced this way read more like human-crafted reports and less like generic AI fluff. Save the prompts now, and you can use them anytime you need a trustworthy, deeply sourced AI write-up.