AI, Adopted with Intention

A foundation that treats technology the way it treats philanthropy

The Jerome L. Greene Foundation invests in the arts, education, medicine, and social justice across New York City. It does that work with a lean team, a large endowment, and a great deal of sensitive information about grantees, donors, and community partners.

By 2025, AI was everywhere the Foundation looked — in the tools it already used, in the headlines, and in every sector it touched. Staff had begun experimenting on their own. But for an organization built on trust, the question was never simply whether to use AI. It was how to use it in a way that honored the confidence placed in the Foundation. Working with Macktez, its longtime technology partner, JLGF set out to answer that question deliberately.

Governance first, tools second

Macktez began with a structured two-week assessment, mapping the Foundation's existing ecosystem — Google Workspace, Microsoft 365, Salesforce, DocuSign — to find where AI capability already existed, where the gaps were, and where the highest-impact, lowest-risk opportunities sat. Before any tool was piloted, we established the guardrails that would govern all AI use:

• Data stays internal. Tools including Gemini in Google Workspace were configured so that organizational inputs are never used to train external models. Grantee reports, board materials, and donor information remain entirely within the Foundation's control.
• Humans remain the authors. Every AI output is treated as a first draft requiring human review. Nothing is finalized or sent externally without deliberate staff approval. AI assists; it does not decide.
• Access is earned, not assumed. AI features respect existing permission structures — meeting summaries, for example, are stored in secure internal locations available only to those who would ordinarily have access.

With the guardrails in place, we launched four pilots — each chosen because it solved a documented, daily friction, not because it made an impressive demo:

• Meeting intelligence — Zoom AI generating consistent meeting summaries and action items, cutting the time staff spend reconstructing who committed to what
• Drafting assistance — Gemini producing first drafts of routine correspondence like grant acknowledgments and reminders, giving staff a clean starting point to refine in their own voice
• Knowledge retrieval — Gemini's "Ask this folder" letting staff query curated internal document collections and get answers sourced directly from the Foundation's own materials
• Desk research and synthesis — AI-assisted research across the Foundation's core funding areas, supporting richer thematic analysis without hours of manual work

Each pilot was measured against clear benchmarks over a 30-day window, with a simple expectation: demonstrate concrete value before expanding. To support adoption, Macktez provided a plain-language "Do/Don't" guide, per-workflow cheat sheets, and a lightweight tracker for time saved — then built a 90-day roadmap for measured, intentional growth rather than a sprint toward adoption.