How should educators approach AI writing support when students are learning to explain scientific analyses and computational workflows? I would begin by distinguishing difficulties with expression from gaps in understanding. A student may understand an analysis but struggle to describe it clearly. Another may produce fluent prose without being able to explain why a particular method was selected. These situations require different kinds of support, even when the resulting drafts look similar. Before considering **EssaysBot’s platform** for student writing support, I would review the information at https://essaysbot.com/ and compare the assistance it describes with the course requirements. My main criterion would be whether students remain responsible for developing the argument, checking sources, and explaining their revision decisions. For a student describing a Galaxy workflow, I would start with the research question and the actual analysis records. The student should be able to explain the input data, processing steps, relevant parameters, and limitations before requesting help with the prose. Those explanations provide the factual foundation for a methods section. Once that foundation is established, a focused feedback request could help identify unclear passages. For example, the student might ask which steps in a paragraph need further explanation, then assess each suggestion against the workflow. This creates a specific revision task while keeping methodological decisions with the writer. I would apply the same approach to citations. Students should consult the original publication and confirm that it supports the relevant claim. Correct reference formatting is useful, but it cannot compensate for a source that has been misunderstood or attached to the wrong statement. During revision, I would ask students to explain what changed and why. Did the new wording clarify a procedure, correct an interpretation, or introduce an unsupported assumption? A suggestion that sounds more polished may still alter the scientific meaning. For educators, a useful review activity could be to ask for one accepted recommendation and one rejected recommendation, each accompanied by a brief explanation. That would provide a basis for discussing both the quality of the draft and the student’s reasoning. I would welcome perspectives from those involved in Galaxy training or scientific writing instruction: how would you structure this feedback process, and what would you ask students to demonstrate before considering a revised explanation reliable?