
Case Study: Grant-Writing GPT
The Problem
Josue, the founder of First Tech Fund, was solely responsible for writing complex, time-intensive grant applications with limited capacity.
He developed a GPT trained on past successful grants and refined it through detailed prompt design.
The Solution
He submitted a complete grant in six hours and secured First Tech Fund’s first-ever AI-supported award.
The Result
Josue and the First Tech Fund Story
There’s a structural inefficiency in how small nonprofits win funding. The grant process demands time, expertise, and context. Josue de Paz, founder of First Tech Fund, had all the context. What he didn’t have was time.
As the sole person responding to grant applications, he faced a constant trade-off between ambition and capacity, as grant applications took him days to complete. So he built a GPT-powered tool, trained on successful grants and designed to produce thoughtful, tailored responses. The process wasn’t seamless and, at first, the model misfired. But Josue did what many miss: he slowed down and invested in the prompt. Specifically, he told the system what it needed to know, what it shouldn’t guess, and when to ask questions.
The real test came when a grant opportunity appeared with a six-hour deadline. Josue applied for the grant with his tool and he won.
The broader point isn’t just about speed. It’s about leverage. With intentional design, AI gave Josue what so many nonprofit leaders need: room to lead.