Researchers moving between literature review, experimental data processing and note keeping need to consolidate research material into reviewable results; the candidate material does not say what material AI receives or which task it completes.
The public material does not describe existing alternatives, so it is unclear who previously did the same task and with what tools.
The candidate material provides no pain description, so it is impossible to confirm which step blocks researchers or what happens if it is not solved.