Converting a Published Paper into a Grant Proposal with Claude

Turning a paper into a grant proposal with AI looks like a copy-paste job and is not one. A paper reports what happened. A proposal argues that something else should happen next and deserves funding. The intellectual substance carries over; the rhetorical direction reverses, and that reversal is the part a language model does not perform on its own.

What transfers and what does not

The literature review, the methods you have validated, the results, and the honest limitations all move across with light editing. What does not transfer cleanly: the gap statement (a paper's gap is closed by the paper, a proposal's gap must still be open), the framing of results (findings in a paper, preliminary data in a proposal), and the innovation claim, which a paper rarely needs to state outright.

Step 1: Extract the gap that is still open

Give the model your paper's introduction and discussion: "From this paper, list every knowledge gap mentioned. For each, mark whether this paper closed it or left it open. For the open ones, phrase each as a one-sentence gap statement suitable for a grant's Specific Aims" (NIAID: Draft Specific Aims).

Human gate: the model over-produces gaps, and several will be gaps your own paper already narrowed. Pick one. A proposal built on one clearly open question beats one hedging across five.

Step 2: Reframe results as preliminary data

Preliminary data has a different job than a results section. It shows the approach works and the team can execute, not that a question is answered. Prompt: "Rewrite this results section as a preliminary data paragraph for a grant. Emphasise feasibility, the direction of the effect, and that the method is already established in the applicant's lab. Do not state that the research question is resolved."

Human gate: check that nothing in the rewrite promises the proposed study's conclusion. A proposal that reads as though the work is already done invites the reviewer to ask why it needs funding.

Step 3: Write the innovation section

This is the section with no equivalent in the paper. Reviewers still expect an explicit statement of what departs from current practice in the concept, the approach, or the methods. Prompt: "Based on this paper and this gap statement, draft an innovation paragraph covering what is new in concept and what is new in method. Label each item as genuine novelty or incremental improvement."

Human gate: delete anything labelled incremental unless it is genuinely central to the proposal. Overclaimed innovation is a routine triage killer, and reviewers who know the field will spot a repackaged increment quickly. The novelty argument here is the same one covered in proving originality in your paper.

Step 4: Generate the budget rationale from methods

The methods section already contains most of what a budget justification needs: personnel effort per procedure, equipment, per-participant costs, and a timeline. Prompt: "From this methods section, list every activity that carries a personnel-time or direct cost, and draft a one-line budget justification for each."

Human gate: the model does not know your institution's rates or salary caps. It produces the structure. The numbers are yours, and they have to reconcile with the effort described in the research plan.

The reframe the model keeps missing

Across all four steps, one correction recurs: the draft explains what was found instead of arguing why the next study is worth funding given what was found. Reviewers are not scoring your past work. They are deciding whether the next question, the approach, and the track record justify the money. Read every paragraph the model returns against that test, and rewrite the ones that still read like a report.

This conversion sits downstream of a finished paper, so the results-to-submission sprint is worth running first, and the broader map of AI's role in grant text is in AI grant writing for academic researchers.

The genre shift from reporting to arguing is the hard part of every section transition, in a paper and in a proposal. The Prompt Pack: Academic Paper Structuring has 25 tested prompts for those transitions across Introduction, Methods, Results, and Discussion, each with a worked example and a warning about where it overreaches.