An honest conversation about craft, trust, and what AI means for nonprofits.
The AI conversation happening in the nonprofit sector right now is mostly noise. The space is full of vendor hype and generic advice repackaged for nonprofits.
Instrumentl did something different.
On May 16, 2026, Instrumentl gathered eight leaders for an intimate dinner in the Bay Area — founders, consultants, frontline development staff, a philanthropy leader, and a fundraising coach with different roles, different org sizes, and different experience with AI. We asked one question and let the conversation flow: what does AI actually mean for this work?
Two hours later, we had more clarity — and more productive disagreement — than we expected.
Here's what the group landed on, and the conversation that got them there.
Use AI to do better work, not just faster work: Sign up for our first ever Instrumentl Course: AI for Grant Work
The participants
Fielding Jezreel — Founder & CEO, Jezreel Consulting; Federal Grants Accelerator
Susan Mernit — Director, Development and Technology, The Early Wealth Partnership; Nonprofit AI Consultant
Mallory Erickson — Founder & CEO, Practivated; Author & Host, What the Fundraising
Cameron Geringer-Pate — Director of Grants and Partnerships, The Hub OC
Danette Fettig Halloran — Corporate & Foundation Relations Manager, The Greater Minneapolis Crisis Nursery; Founder, The Empowered Person – AI
Michelle Maryns — Founder & CEO, We Sparkle Co.; President, Minnesota Council on Foundations
The five things the group agrees on
1. Quality expectations are going up, not down. When writing becomes cheap and ubiquitous, the things that can't be automated — judgment, relationships, authentic voice — become more valuable. AI doesn't lower the bar. It raises it.
2. Expertise is the prerequisite. The practitioners who use AI most effectively are the ones with the strongest opinions about what great looks like. You can't delegate taste. AI amplifies what grant professionals bring, which makes knowing your craft deeply more important now, not less.
3. The grant profession is evolving to have even more leverage. The sentence-level work is becoming automatable. The strategic work of connecting dots, curating what a funder needs to see, making sense of data, and stewarding relationships is becoming more central.
4. The real opportunity isn't in the proposal. It's in the workflow around it. The most interesting applications of AI for nonprofit work aren't about writing faster. They're about changing how information flows from program teams to grant teams to funders in ways that build more authentic relationships.
5. The funder-grantee conversation is overdue. The sector is in a gray zone where some funders ban AI, others screen for it, and many haven't formed a position. Navigating the future well will require transparency and shared agreement that have historically been a challenge.
A deeper look at the conversation
On craft: what changes, what doesn't
The group moved past the anxiety conversation quickly — is AI writing good enough, can funders tell — toward something more useful. What does craft actually mean when the executional part of writing gets easier?
Fielding reframed it early. The best grant writers have always been formulaic, working with patterns, structures, and repeatable architecture. That's exactly what AI is good at. Train it on your formula, and you free yourself for the work that can't be templated — the strategic thinking and judgment calls that no prompt can make for you.
The room built on that. Mallory connected craft to trust, arguing that what we've historically called craft is giving way to something harder to fake: how you connect dots and earn credibility with a funder over time. Michelle named the function that stays irreplaceable regardless of how good the tools get: sense-making. The human who can take complex information and turn it into a story that lands.
What the group agreed on is that this shift raises the stakes for developing real expertise, not lowers them. Susan, who has spent years coaching nonprofits on AI adoption, was direct about it. Getting good at AI is like learning a musical instrument. The competency compounds, but only if you treat it as a practice — something you commit to, invest in upfront, and keep building.
On the profession: from writer to something more
If not "grant writer" in 2029, then what?
Fielding put the underlying tension plainly. The proposal has always been downstream of the real work of knowing an organization deeply, understanding what a funder actually cares about, and curating what belongs in front of whom and when. Writing is the last step, and increasingly, the most automatable one. What's becoming more central is the curatorial judgment that precedes it.
Susan pushed toward "funding strategist" as the frame that felt most accurate. Someone who thinks about mission, revenue, and opportunity simultaneously and builds toward a funding picture rather than just responding to open RFPs.
What the group was ultimately optimistic about is that AI clears the way for that bigger role. Mallory made the point directly: a lot of the people drawn to this work got here because they care about a mission, not because they love administrative complexity. The more AI can hold the operational burden, the more the people doing this work get to do what actually drew them to it — the relationship-building, the strategic thinking, the advocacy that makes the difference.
On workflow: where the real shift is happening
At Cameron’s organization, which has grown up with AI already part of its grants process, the shift isn't in the proposals, it's in the conversations that feed them. Instead of frantically transcribing during program check-ins to collect data, his team just talks — about the causes they serve and what's being learned. AI helps capture it while the conversation stays human. The result, as he put it, is that 80% of the work becomes relational.
Fielding extended that thread into a meaningful workflow inversion. What if the funder relationship stopped being reactive, triggered by deadlines and reporting requirements, and became proactive instead? One conversation with your program team, well-captured, becomes the source material for every funder communication, each shaped to what that particular funder cares about. It’s more authentic, not less, because the source is a real conversation rather than a document built backward from a requirement.
Danette pushed the vision further, toward an organizational "learning layer" that doesn't just inform grant writing but changes how nonprofits understand themselves in real time. Staff input, community feedback, and program data, all captured and surfaced in ways that let organizations show their impact beyond a polished proposal; the intangibles like good judgment, the ability to pivot, and the capacity to respond.
Mallory grounded it in something most grant professionals recognize immediately: the motivational reality of compliance. The urgency before a deadline is real. The attention to post-award tracking, nine months later, is not. The opportunity is to build systems that account for how humans actually work, handling the parts that drain motivation so the people doing this work can lean into the parts that fuel it.
On funders: navigating the gray zone
The sector is in the middle of something, and the group didn't pretend otherwise.
Michelle, who sees the landscape from the philanthropy side, walked through the range of funders who ban AI-written proposals entirely, funders who welcome it as an equalizer for under-resourced organizations, and funders using AI themselves just to manage the volume of applications they're now receiving. There's no consensus, and no clear timeline for one.
What made the conversation productive is that the group didn't stop at the problem. Fielding pointed toward the posture that actually moves things forward. That is, pushing for funder-grantee dialogue built around shared goals rather than rules handed down from one side. The power differential is real, she acknowledged. But the more both sides can sit at the table asking what actually makes sense to achieve a common outcome, the better the answer will be for everyone.
Mallory held both sides of the trust question clearly. Using AI carelessly in external-facing work is a genuine risk in a sector where donor trust is the currency. And opting out entirely, when AI could help an organization use donor dollars more effectively, carries its own responsibility. The path forward isn't a policy. It's an ongoing conversation with funders, communities, and each other.
Cameron pointed to where the sector's energy should go. The organizations with strong, trusted funder relationships are the ones positioned to test what's possible and bring the learning back, not just for their organization, but for the field.
Lead your organization through change: Sign up for our first ever Instrumentl Course: AI for Grant Work
The bigger picture
The administrative complexity that consumes so much of a grant team's time has always been in tension with the actual purpose to build the relationships, strategy, and understanding of what a community needs. Then, connecting the right resources to it.
Michelle ended the dinner with a vision that stayed with the room. She thinks about mycelium, the underground networks through which trees communicate needs and share resources in real time, without any single point of control. What if we could build something like that for the sector? AI that knows, in real time, where the needs are. Resources flow accordingly, and there’s less need for the proxies, like proposals, reports, and backward-looking documentation, we've always relied on in their absence.
Watch the most honest conversation the nonprofit sector is having about AI.
