The Allen Institute for AI (AI2), in collaboration with researchers at the University of Pennsylvania, has unveiled LitPivot, an AI-powered system designed to guide researchers through the process of developing and refining research ideas in relation to existing academic literature. The tool addresses a common challenge in academic research: navigating how new ideas relate to prior work, a process researchers call "literature-initiated pivots." LitPivot helps researchers understand whether their ideas are truly novel, identify gaps in existing research they could address, or find ways to build upon established work. LitPivot works by allowing researchers to write short research proposals in an editor and segment their ideas into distinct facets—the problem being addressed, the proposed contribution, and the evaluation plan. When researchers select a particular facet, the system automatically groups related papers and highlights the most relevant ones. Based on papers the researcher selects, LitPivot generates AI-powered critiques supported by direct excerpts from academic papers, along with specific suggestions for how to revise the proposal. Researchers retain full control, able to accept, edit, or reject any suggestions and request new literature searches as their ideas evolve. A validation study with 17 researchers demonstrated the effectiveness of LitPivot's approach. When compared to researchers developing ideas using a traditional text editor paired with an AI chat tool for paper analysis, those using LitPivot produced research proposals that domain experts rated as significantly better grounded in the literature. Participants also reported feeling they had a stronger understanding of the surrounding literature and perceived their revised ideas as more novel and useful. LitPivot will be presented at the UIST 2026 conference in Detroit.