Every veteran nonprofit I talk to is asking the same question right now: should we be using AI? The honest answer is yes, in some places, and absolutely not in others. The line between those places isn't about technology. It's about trust.
Your organization runs on trust. A veteran who calls your line, a widow filling out an intake form, a donor deciding whether to give again: each one is deciding whether you're the kind of organization that shows up. AI can help you show up more often. It can also make you look like you didn't show up at all.
Start with the outcome, not the tool
The mistake I see most often is starting with a tool. Someone on the board saw a demo, or a vendor offered a discount, and suddenly the team is "doing AI." Six months later, nobody can say what it changed.
Flip the order. Write down the outcome first: answer every inquiry within one business day, cut grant-report prep from a week to two days, get the gala recap out while people still care. Then ask whether AI is the fastest, safest way to get there. Sometimes it is. Sometimes the answer is a better form or a shared calendar.
Where AI earns its place
For a small team, AI is most useful in the work nobody sees:
- First drafts. Newsletters, social posts, grant narratives, and event recaps. A person still edits, approves, and owns every word, but nobody starts from a blank page.
- Organizing what you already have. Summarizing survey responses, tagging donor notes, pulling themes out of a stack of program feedback.
- Routing. Sorting incoming inquiries so a benefits question reaches the right person instead of sitting in a general inbox.
- Repurposing. Turning one strong speech or report into a month of content across channels.
- Reporting. Pulling numbers from your CRM, email platform, and social accounts into one view your board can read.
Notice the pattern. In every case, AI is giving time back to the people on your team so they can spend it on the people you serve.
Where a person has to show up
Some moments should never be handed to a machine, no matter how good it sounds:
- Crisis. Any message that suggests a veteran is in distress goes to a trained person immediately. Your systems should be built to flag it, not answer it.
- First contact with a family. The first time a spouse, parent, or survivor reaches out, they should hear from a human being.
- Stewardship of major donors. A thank-you that sounds generated tells a donor exactly how much you value them.
- Advocacy and public statements. Your organization's voice on policy, legislation, or litigation carries legal and reputational weight. A person writes it, and a person signs off on it.
- Anything involving personal information. Service records, medical details, and benefits information don't go into tools your organization hasn't vetted and approved.
Put guardrails in writing
You don't need a 40-page policy. You need one page that every staff member and volunteer has read. It should answer four questions: Which tools are approved? What information can never go into them? Who reviews AI-assisted content before it's published? And who do people ask when they're not sure?
Write it down, share it, and revisit it every quarter. The tools change fast. Your standards shouldn't.
Measure it like anything else
AI is an investment of money and attention, so treat it that way. Before you start, write down what you expect it to change and how you'll know. After 90 days, look at the numbers. If it saved real hours or improved response times, keep it. If it didn't, stop paying for it.
That's the same standard we apply to every channel and every campaign: measure before you spend, and keep only what earns its place.
