Walton Family Foundation

The situation

The Walton Family Foundation’s Environment Program funds thousands of nature-based projects across the United States to reduce climate risk and improve river and ocean health.

The facts about those projects were locked in narrative reports and PDFs written by dozens of partner organizations, each in its own format. The Foundation could not answer portfolio-level questions, such as “show a map of all projects in Louisiana”, without asking grantees to change how they report.

What we built

• AI-assisted extraction. Large language models read the backlog of narrative reports and pulled out the core facts: project identifiers, locations, grant amounts and projected environmental outcomes such as acres restored.
• A central data warehouse. The extracted data flows into one place, with interactive dashboards on top.

Grantees kept reporting the way they always had.

The result

Program staff can see the geographic spread of the portfolio, analyze how it lines up with federal funding, and total up ecological benefits across projects, all from information grantees were already providing.

The takeaway

Narrative reporting is a store of institutional knowledge. Varied formats are not a data quality problem to fix with rigid new software. AI can read the reports as they are and meet people where they already work.