Data Management Plans (DMPs) are required by DFG, ERC, Horizon Europe, NIH, and most major research funders. They’re formulaic documents with specific required sections — exactly the type of document AI helps produce efficiently.
What a DMP Requires
A standard research DMP covers: data description (what data you’ll produce, formats, volume), data collection methods, metadata and documentation standards, data sharing policy (what will be shared, when, where), long-term preservation plans, and data security measures. The DFG guidelines (DFG Leitlinien zum Umgang mit Forschungsdaten) specify these requirements in detail; so does the Horizon Europe DMP template.
Using AI to Draft Your DMP
Download the specific DMP template for your funder. Paste the template requirements into Claude with your project description: “I am applying for a DFG grant for a project on [describe it]. The project will produce [data types]. Here are the DMP requirements [paste them]. Draft a DMP addressing each requirement.” The output will be a complete draft covering all sections. Your task is then to verify factual claims (does your institution actually have the data repository you claim? are the metadata standards you mention appropriate for your field?).
Field-Specific Metadata Standards
Different fields have different metadata standards — Darwin Core for biological specimens, MIAME for microarray data, DDI for social science survey data, FITS for astronomy. Ask Claude: “What are the standard metadata schemas for [your research type]?” before writing the metadata section. Funders check whether you’ve named field-appropriate standards, not just generic ones.
Repository Selection
Where you plan to deposit data matters. Domain-specific repositories (GenBank for genomics, PANGAEA for earth science, OSF for social science) are preferred over generic ones (Zenodo, Figshare) when they exist. Ask Claude: “What are the standard data repositories for [your field] that would satisfy DFG/Horizon Europe requirements?” Then verify that those repositories are still active and accepting your data type.
FAIR Data Principles
FAIR data (Findable, Accessible, Interoperable, Reusable) is the framework most funders use to evaluate DMPs. Ask Claude to explain what FAIR means specifically for your project type, then use that framing in your DMP to signal to reviewers that you understand the principles — not just checked the boxes.
