There is something interesting happening in nonprofit circles right now around AI. Leaders are exploring it, piloting it, and in many cases already using it, while at the same time carrying a quiet unease about where it all leads. Not fear exactly, more an open question about what role people play in a world where so much can be automated.
AI is a tool. That sounds almost too simple given how much has been written about it, but I find it’s the most useful place to start. Because once you accept that framing, a more important question emerges: a tool intended for what, and are we using it in a way that serves our mission? For nonprofits whose missions are rooted in human connection, that question deserves a real answer.
The Jarvis vs. Ultron Problem
I find myself coming back to a somewhat unlikely reference point for this. In the Marvel universe, Tony Stark builds two very different kinds of AI. Jarvis is designed to support him — handling complexity, surfacing information, and amplifying what he’s already capable of doing. Ultron is given an open-ended objective and left to pursue it without meaningful human oversight. One makes Stark more effective. The other nearly ends the world.
The difference isn’t the technology. It’s the design philosophy behind it, and the role humans play in keeping it pointed in the right direction.
For nonprofit organizations, this distinction matters more than almost any other type of organization. Mission driven work is fundamentally about human connection. The donor who gives because they believe in your mission. The community member who shows up because they feel seen and supported. The staff member who stays because the work feels meaningful. That connection has nothing to do with AI; it is about how you show up for your communities.
What AI Can and Cannot Do
At AGP, we use AI to help build solutions that help organizations connect to their communities and support their mission. We have used it to synthesize and analyze data, translate content across languages, and build AI search experiences that make it easier for people to find what they need. These are just some examples of AI capabilities that free up real time and capacity for the people doing this work.
Remember, AI is just a tool. Deciding whether that appeal is true to your organization’s voice and values still belongs to a person. The stories that you share belong to your communities. The emotions you convey to help drive impact are human. Building the kind of trust that makes a community feel genuinely served will always come from the people behind the mission. That is why every AI solution we build involves real human review and judgment, because the judgment, context, and empathy that people bring is the part that actually makes the work meaningful.
The best AI implementations I’ve seen in the nonprofit sector tend to do the same thing: use AI to remove friction and free up human capacity, then direct that capacity back toward the work that requires judgment, empathy, and relationship.
The Real Risk
I worry about using AI without intention, treating outputs as finished work, automating decisions that deserve human review, adopting tools because everyone else seems to be, or thinking you can replace jobs with AI to save on your bottom line. At AGP, we know that AI is just another piece of technology that we will use to help better serve our clients and their mission.
For mission-driven organizations, AI should serve your mission, your communities, and your people. It should make your people more effective, not make them feel unnecessary. And above all, it should amplify the human connection at the center of your work.
Jarvis, not Ultron. The technology is the same either way. The choice is in how you use it.
