In Our Toolkit
I am not interested in using AI to produce more content. I am interested in using it to produce better work. In fundraising, the writing has consequences, so I treat AI like a collaborator that needs direction, not a machine that gets to run on autopilot.
Before we write, I set a few non-negotiables. These are such important guardrails in my work that they have become rules for all the AI writing I do.
- Show your work. No numbers or claims without sources. If we cannot source it, we remove it or mark it as a placeholder.
- Avoid performative social-media cadence. I want complete sentences and real paragraphs. I do not want rhetorical fragments, filler transitions, or writing that looks designed to be skimmed instead of understood.
- Sound like me. I will not send language that reads like it came from a generic template.
- Wow me. “Nice” is not the bar. This is serious fundraising. The writing has to earn belief and feel worthy of the investment.
My workflow starts by slowing the machine down on purpose. I almost always begin with one instruction: Do not start writing yet. Drafting too early is how you get confident prose with weak logic. I would rather take two minutes to lock the brief than spend an hour cleaning up a draft that never had a clear target.
Next, I make the tool mirror back what it thinks I am asking for. Before we write a single paragraph, I ask for a repeat-back: the goal, the audience, what the proposal or grant is about, etc. funder is actually asking for, etc. That repeat-back step is where my own thinking gets sharper, because it shows me immediately what is fuzzy or misaligned.
Only then do we outline. We build the structure first: the argument, the sections, and the few things that must be true for the piece to be fundable. After that, we write one section at a time. We start with the opening and get it right before we move on. The rule is simple: we finish each section before we earn the right to move to the next one.
Then we iterate until it lands. That is my secret sauce, and it is why we produce strong proposals and reports. I do not accept “good enough,” and I never settle for the first draft. AI does not make the writing process easy, but it does make it possible to revise more, and revise faster, which is how the work gets better.
At the end, I run a different kind of check. I ask for a whole-document critique, and I ask for it to be done like a skeptical program officer would read it. This is also where I will often say, “It sounds nice, but it is not doing it for me yet.” That is not pickiness. That is respect for the work. If we are asking for significant, serious philanthropic investment, the writing needs to read like it.