How Patrick J. McGovern Foundation Uses Claude to Standardize Grant Evaluation

The Patrick J. McGovern Foundation occupies an unusual position in philanthropy: a mission-driven organization that simultaneously funds AI and data initiatives and runs its own technology team capable of building internal tools. That combination gave PJMF both the motivation and the means to address a problem that affects the entire grant-making sector—the absence of standardized financial due diligence in the evaluation of nonprofit grant applicants. Most foundations rely on staff without specialized financial training to assess complex documents that arrive in inconsistent formats, creating variability that disadvantages both funders and applicants.

PJMF’s pre-AI process was built on Excel spreadsheets where staff extracted specific financial metrics by hand. The system worked until it didn’t—every edge case, every document with an unusual date cutoff or non-standard format, forced an individual judgment call. Staff would annotate their work with explanations of adjustments they made, but the next reviewer wouldn’t necessarily make the same decision. The result was an evaluation process that was accurate in aggregate but inconsistent in execution, undermining confidence in comparisons across applications.

PJMF evaluated multiple AI models before selecting Claude, with data privacy emerging as the decisive factor. Grant evaluation involves sensitive nonprofit financial records; the foundation required an AI partner whose data handling practices matched its commitment to responsible AI. Grant Guardian uses Claude to ingest nonprofit financial documents regardless of format, extract key metrics, and organize them into standardized profiles aligned with each funder’s specific evaluation criteria. Funders can configure their evaluation parameters—for example, adjusting risk thresholds for newer organizations with shorter financial histories—without penalizing valid applicants.

The impact on PJMF’s internal workflow was immediate and measurable. What previously required staff training, manual extraction, and individual judgment calls now produces consistent analysis applied uniformly across every application. United Way affiliates using Grant Guardian report significant time savings and the ability to manage their entire due diligence portfolio internally, without engaging outside volunteers or consultants. The tool does not generate a single composite score; instead it surfaces areas where deeper conversation with the nonprofit is warranted, keeping human judgment at the center of funding decisions.

PJMF’s broader vision extends beyond operational efficiency. The foundation sees AI transforming philanthropy along two dimensions: standardizing internal processes so foundations of all sizes can apply rigorous, consistent evaluation, and helping grantees themselves leverage AI for mission-driven work. By building and sharing Grant Guardian across 100-plus philanthropies, PJMF is establishing a model for how AI can democratize access to sophisticated analytical capabilities in the social sector—turning what was once a resource advantage for large foundations into a tool available to any organization willing to engage with it.