Nonprofit leaders around the world suggest that artificial intelligence is improving the efficiency of their offices, but without adding more capacity. How did we get to where AI has gradually crept into the processes of the nonprofit workforce?

AI has begun to make its way into the day-to-day work of many nonprofits. But it’s often not through formal strategies or organisational decisions. Instead, staff are making small, practical choices in isolation.

Maybe someone uses it to draft a proposal more quickly. Someone else relies on it to summarise a long report before a meeting. A communications team experiments with it for social media content. These small adjustments make the work a bit easier and faster.

For organisations that are often under-resourced and stretched thin, this is not trivial. Being able to save time on routine tasks or produce materials more efficiently can make a real difference, showing the impact of AI gradually making its way into the nonprofit workforce.

But these changes are happening in an unstructured way. In most organisations, there has not been a broader conversation about what role AI should play or how it should be used. There are few internal guidelines, limited training, and little clarity on questions like data use, verification, or where to draw the line between support and substitution. In the third sector, discussing AI replacing roles is often taboo, turning workplace efficiency discussions into moral or ethical debates. This drags on limited budgets and adds pressure to already stretched resources.

In practice, this means that individuals with a vocation to use AI in the workplace, rather than organisations, largely drive adoption.

The reflections in this piece are based on an ongoing personal research effort. I hope to publish the findings at some point, but for now, it is a conversation with fellow executive directors from my past roles. Over the past months, I have spoken with organisations across regions and contexts—10 from Latin America, five from the United States, five from Europe, seven from Africa, and five from Asia, covering a wide range of funding levels and operational capacities. As mentioned, this personal endeavour is still unpublished, and what follows should be read as early observations from those conversations rather than definitive conclusions as AI gradually makes its way into the nonprofit workforce.

Read the full article about how AI made its way into the nonprofit workforce by Gastón Wright at Alliance Magazine.