Mission-driven organizations are under pressure to “do something with AI.” Funders ask about it. Boards ask about it. But most of the AI initiatives launched in this sector fail — not because the technology doesn’t work, but because the use case was wrong from the start.
This is a framework for deciding where AI belongs in your portfolio, and where it does not.
Start with the decision, not the technology
The question is never “should we use AI?” It’s “what decision are we struggling to make well, and would better evidence change it?” If the answer is no, no amount of machine learning will help.
(This is a migrated stub — paste the full essay body here. Everything below the frontmatter is standard Markdown: headings, lists, links, blockquotes, code.)
What to fund
- Problems where you already have the data, and the bottleneck is analysis capacity.
- Decisions made repeatedly, where consistency matters more than nuance.
- Reporting burdens that consume senior time better spent on programs.
What to skip
- Anything where the training data would encode a bias you can’t defend to your community.
- Novelty projects funded because a grant mentioned AI.
- Systems your team can’t run after the consultant leaves.