Agree the remaining question before pricing the pilot: presentation, seating, handling, or the whole workflow. Ask your colleague to define and price the work needed for that question, including your assistance, then agree a stopping point. A small study can be approved without committing to the arm or quietly expanding into a cell build.
That is a useful boundary. Keep the study's observations distinct from the full-cell cost and feasibility questions, especially if one favourable sample tempts either of you to call the whole proposal settled.