From 90 applications to 200, without losing rigor

# Public & Social Impact

# ChatGPT

# Awareness

# Personal

How a Texas foundation president used GPT-5 to sharpen grant review and strengthen mission language

The foundation Kim Davis runs typically reviews about 90 grant applications per cycle. When more than 200 came in this round, he turned to ChatGPT-5 to help decide which to fund. The model didn’t just throw out an answer. Instead, it explained why certain of the competing mission statements worked better, and then, without being asked, offered alternative versions tailored for different goals. One was bold and brand-forward, the kind of statement Davis said board members at the Carl B. and Florence E. King Foundation could easily remember. Another was warm and approachable, better suited for a public-facing website. Each came with a clear rationale, helping Davis – the foundation’s president – quickly weigh the trade-offs and choose the best fit. In the past, Davis sometimes used earlier versions of ChatGPT to speed up the reviews, but those models could wander from instructions, produce inconsistent results, or require repeated prompts to stay on track. GPT-5, by contrast, delivered faster, more precise, and more context-aware support right from the start. “It’s like having a partner who works alongside you," Davis says. “It doesn’t just give me an answer, it tells me why it’s the right one, and shows me other paths I could take.” Davis was among the nonprofit leaders who participated in OpenAI’s July 2025 Nonprofit Jam, which provided hands-on training in how to use AI to save time, expand reach, and better serve communities. With $80 to $100 million in endowment assets and $3.3 million in annual grants, the King Foundation is investing in communities across Texas and beyond. And by using GPT-5, Davis and its leadership team are communicating more clearly, making stronger impressions with stakeholders, and improving their most valuable partnerships.