Currently open: accepting 1–2 engagements for the second half of 2026 (July – December). This page will update when the roster is filled.
One or two engagements per year go to nonprofit organizations working on AI-adjacent problems. If that describes your work and you can't hire a firm at market rates, reach out.
Who this is for
The invitation is scoped intentionally. It fits organizations where the AI question is real and the budget for outside advisory isn't. Specifically:
- Nonprofits building AI capabilities under nonprofit constraints. Program teams trying to responsibly use LLMs for workflow, program evaluation, or member services; leadership trying to build an AI use policy, a model-governance approach, or an eval discipline without a for-profit-sized budget.
- Journalism and civic-information organizations addressing AI misinformation. Newsrooms, fact-checking outfits, and civic-tech projects that need help thinking through detection tooling, editorial policy, or content-provenance systems.
- Workforce and education programs adapting to AI displacement. Community colleges, workforce boards, apprenticeship programs, and skills-training nonprofits working out what changes when the entry-level jobs their trainees were preparing for start absorbing AI.
Other AI-adjacent missions with a plausible fit are welcome to ask. The mission areas above are illustrative, not exhaustive.
What the engagement looks like
Pro bono engagements are structured the same way paid ones are, at proportional scale. Typically:
- Scoping call. Sixty to ninety minutes to understand the specific problem, the constraints, and what "done" would look like. No prep required.
- A defined deliverable. Usually a written artifact — a decision memo, an eval-framework plan, a governance draft, a vendor-selection read — rather than open-ended advisory hours. Pro bono time is scarce; a defined output uses it well.
- A time budget, agreed up front. Most engagements fit in twenty to forty hours over four to eight weeks. Larger commitments are rare and only in cases where the counterfactual is the work not getting done at all.
- No public attribution required. If you want to credit the practice publicly, that's welcome. If your board or funders would rather the engagement stay private, that's fine too.
What this is not
To spare us both a mismatched intake call: this is not a general nonprofit-services offer. If you have a well-funded development office and could reasonably retain a firm at market rates, this program isn't the right fit — the annual capacity is small and the intent is to reach organizations for whom paid advisory is genuinely out of reach.
It is also not a bench of unpaid labor. What I can do is bring the same practice discipline that a paid engagement gets — evals, governance, program design, vendor read — to a well-scoped problem the organization couldn't otherwise fund. What I cannot do is take on ongoing operational responsibility, staff a program, or build software.
Reach out
Send a short note describing what your organization does, the specific AI question you're trying to answer, and what a good outcome would look like from an engagement. If it's a fit, we'll set a scoping call. If it's not, I'll say so and try to point you at something that is.