“Where does the organisation stand on AI, and what should be addressed first?”
A board member, a donor questionnaire or your own staff have asked where the organisation stands on AI. What exists is an impression: some people use it, nobody is certain who, and the tools in use have never been listed.
Two sector findings explain why the impression is usually wrong in the same direction. In a benchmark study of 346 nonprofits, 92% report using AI while 81% use it individually, without shared workflows source. Surveying 215 foundations and 451 nonprofits in April 2025, the Center for Effective Philanthropy reports that almost two-thirds say none or just a few of their staff have a solid understanding of AI source.
Each dimension carries a score, the finding behind it, and the action it implies. Four illustrations follow.
Illustrative finding: three of five departments have entered beneficiary information into a consumer tool with no data processing agreement; two have not. The action is a data classification table and an approved tool list, both inside the first thirty days.
Illustrative finding: fundraising uses AI weekly, finance not at all, and neither has had training built on its own tasks. The action is one general session and two department labs, with fundraising first.
A table of every AI tool in use: name, department, purpose, account type (organisation or personal), the data class it touches, and whether it is approved. This is the artefact funders ask for most often and the one least often ready.
Ranked, not exhaustive. A typical first three: publish the approved tool list; run the general session; draft the clause covering beneficiary data. Each with an owner and an effort estimate.
They share the same framework and differ in depth, in who answers, and in what the result can be used for. Two are free and self-administered on this site.
Typical duration: Two to three weeks
Priced on organisation type and size, and fixed in writing after a 30-minute call. The assessment is usually the first component of a package; the estimator shows what adding a policy or training does to the total.
If a different one fits better, or none does yet, we will say so on the call.