Preparing a research budget is complex and time-consuming because researchers must estimate personnel, equipment, software, materials, travel, and other costs while complying with university policies, funding requirements, and procurement procedures. We propose an AI Research Budget Planner that allows researchers to describe their project and automatically generates a draft budget with recommended cost categories, reference prices, explanations, and policy checks.

The system can combine LLMs with university regulations, procurement guidelines, historical budgets, and reference pricing through a RAG-based knowledge layer. It is feasible to prototype using existing policies and sample project budgets without training a new model. The next step is to pilot the solution with a small set of research proposals, validate its recommendations with Finance and research management teams, and progressively integrate real procurement and historical cost data.