Forming interdisciplinary research teams and crafting competitive research plans is a highly customized, time‐intensive, and often fragmented process that can detract from researchers’ time and passion in making innovation. We present ResearchConnect, an AI‐powered web platform that automates researcher profiling, intelligent team formation, and research ideation. The platform takes multiple forms of input (PDFs, manual text, or website links) to extract research‐specific keywords and cluster researchers into thematically coherent teams. It adopts various large language models (LLMs) to suggest possible interdisciplinary research topics based on each team's collective expertise. To evaluate its performance, we used datasets extracted from 250 NSF‐funded publicly accessible research summaries, reconstructing researcher profiles and testing our automated team formation algorithm against original project groupings. Furthermore, we assessed ideation relevancy by comparing AI‐generated ideation summaries across different LLMs (GPT, Grok, LLaMA) against expert‐written summaries using semantic similarity metrics. The results demonstrate that ResearchConnect effectively replicates human‐crafted summaries, underscoring its potential to streamline early‐stage scientific brainstorming, collaboration, and ideation.
Jadhav et al. (Wed,) studied this question.