Our Commitment
Every choice we make about AI comes back to the same thing the rest of our work does: helping charities and social-good organizations raise more money, reach more people, and get on with the mission. We just do it with people in the loop.
We are not here to hype AI, and we are not hiding from it. We use it carefully, in service of the work, with people firmly in charge.
What Clients Can Count On
- People are always responsible: AI can speed up a draft, sort through data or spot a pattern, but it never gets the final say. Anything that reaches you has been read and signed off by someone accountable for it, because the result is ours, not the tool’s.
- We tell you when we’ve used it: If AI has meaningfully shaped something we hand you, we can say so and explain where.
- You decide how much AI is in your work: Some organizations want us bringing AI ideas to the table, others would rather we kept it out of the way, and both are completely fine. You can set the level you’re comfortable with, and you can change it whenever you like.
- Your supporters’ data stays out of it: Donor names, contact details and giving history never go into AI tools. We work from anonymized, aggregated or already-public information, and anything sensitive stays where it belongs, in the systems built to hold it.
- Nothing goes out unchecked: AI can be wrong and still sound certain, so everything is checked by a person before it reaches you. We verify the facts, watch for bias, and make sure it sounds like your organization rather than a machine. If we’re ever unsure about something, we’ll tell you.
- We keep up, and we keep checking: The tools move quickly, so we keep testing what’s worth using, set aside anything that stops meeting our standards, and revisit how we work as the ground shifts. What we’ve promised you here holds whatever the tools do next.
The Tools We Use Today
We maintain a known set of tools so our information stays in shared, central places instead of scattered across personal accounts:
- ClickUp: Brain and Super Agents; runs ChatGPT, Gemini, or Claude underneath.
- Google Gemini: Gemini Pro via Workspace, connected to our Drive.
- Claude: Team plan including Claude Code and Cowork.
- AI Note-takers: In meetings, including client calls.
- Figma Make: Design and prototyping.
“No supporter, donor, or constituent personal data in any AI tool.”
Policy Guardrails
The Hard Lines (Never)
- No supporter, donor, or constituent personal data in any AI tool.
- Nothing published or sent without a human reviewing and verifying it.
- No profiling or scoring donors across privacy lines or sector ethics.
- No misleading, biased, or discriminatory content.
- Nothing that breaks a contract, platform terms, or the law.
- No mixing personal and work accounts (these tools remember).
Where AI Helps (Examples)
- Research and synthesis; pulling threads together.
- Drafting and ideation: outlines, first drafts, subject lines.
- Quality checks: grammar, clarity, rewriting, QA.
- Coding and QA support: snippets, templates, debugging.
- Summarizing documents, transcripts, and long threads.
- Speeding up internal workflow.