Nonprofit AI case studies
Does AI actually increase donations? Here is the documented evidence.
In documented cases, yes. Save the Children Australia lifted appeal response rates 18 percent with predictive upgrade modeling. Make-A-Wish Arizona raised $3.3 million, 122 percent of goal, at its 2025 Wish Ball. Disaster-recovery nonprofit SBP saw a 59 percent increase in donor conversion from AI-suggested ask amounts. Details below, with the organization, the tool, and the method named in each case.
All twelve case studies are documented in full in Artificial Intelligence for Nonprofit Fundraising by Dale Nirvani Pfeifer.
Save the Children Australia
Save the Children Australia: an 18% response lift from predictive upgrade modeling
- Tool: Dataro
- Use case: donor upgrades
- Result: +18% response rate vs. traditional segmentation
The problem.
Recurring donors provide stability, but growth depends on knowing which of them are ready to give more. Traditional segmentation by recency, frequency, and gift size describes past behavior; it does not reliably predict who is ready to upgrade.
The method.
Save the Children Australia partnered with Dataro to apply predictive modeling to its donor database. Trained on historical giving behavior and engagement data, the model scored supporters on upgrade readiness, and outreach went to the highest-ranked segments. Control-group testing validated performance, keeping a human decision between the score and the donor.
The result.
According to Dataro’s published case materials, the model generated an 18 percent increase in response rate compared to the organization’s traditional RFV segmentation, with more gifts from fewer mail pieces sent. The gain came from deciding who to ask, and when, not from asking more.
Documented in Chapter 7 of the book, alongside Dataro’s published case study.
Make-A-Wish Arizona
Make-A-Wish Arizona: a record $3.3 million Wish Ball on one platform
- Tool: OneCause (Bonterra)
- Use case: event fundraising
- Result: $3.3M raised, 122% of goal
The problem.
The Wish Ball had grown into a signature event drawing nearly one thousand guests, and complexity grew faster than the team: long check-in lines, fragmented data across sponsorships and auctions, manual item procurement, and a two-week data cleanup before stewardship could begin.
The method.
The team consolidated the full event lifecycle on the OneCause Fundraising Platform: auction items and bidding, sponsorship sales and invoicing, QR-code check-in with preregistration, and every supporter’s activity in a single record. Donor data stayed live in the platform, so post-event cultivation could start Monday morning.
The result.
The 2025 Wish Ball raised $3.3 million, 122 percent of the event goal. Auctions generated $526,000, the paddle raise drove $2 million toward Fund-A-Wish, pre-event bidding reached $80,000 before guests arrived, and check-in time dropped by 30 minutes. The organization has since expanded the platform to nearly all of its events.
Documented in Chapter 11 of the book, alongside OneCause’s published case study (2025).
Center for Victims of Torture
Center for Victims of Torture: 150+ grants, one visible pipeline
- Tool: Instrumentl
- Use case: grants management
- Result: new general operating funds
The problem.
More than 150 grants a year managed across spreadsheets, with deadlines living in individual staff members' heads. Grant work was reactive: the next deadline set the agenda, and prospecting for new funders happened only when someone found spare hours.
The method.
CVT moved its grants operation into Instrumentl: deadlines, tasks, documents, and funder research in one shared workspace, with historical data migrated for continuity. The team used structured filters and peer prospecting to surface aligned funders in new regions, and Instrumentl Apply to streamline application drafting. The team’s judgment still decided which opportunities to pursue; the system made the whole pipeline visible.
The result.
Grant work shifted from reactive deadline management to a managed pipeline. CVT expanded outreach into new geographic regions, pursued quick-turnaround opportunities it previously had to skip, and secured general operating funds in a new regional program: the hardest kind of money to raise and the most valuable kind to have.
Documented in Chapter 8 of the book.
And nine more
Nine more, documented in the book.
Each pairs a working framework with a named organization that tested it. Full write-ups will publish here; until then, the complete versions live in the book.
Tarjimly
ethical AI translation at scale
AI first-pass translation with human volunteers in the loop. Language assistance for more than 609,000 refugees in 2024.
Chapter 3
Island Senior Resources
finding hidden capacity
DonorSearch by EverTrue surfaced planned-giving and major-gift prospects a small human-services team had no way to spot.
Chapter 4
Butler University
video personalization at scale
Gravyty personalized video turned Giving Day outreach from a blast into a conversation.
Chapter 6
Foundation Fighting Blindness
one view of forty years of data
Virtuous Analytics replaced a fragmented legacy stack with real-time insight teams could act on.
Chapter 9
Joe Nuxhall Miracle League Fields
racing a $1M match
Bloomerang’s Penny helped a tiny team run personal outreach fast enough to capture a one-million-dollar matching gift.
Chapter 10
Community Rebuilders
AI in the daily workflow
Microsoft 365 Copilot embedded in everyday tools, saving roughly ten hours of staff time a week.
Chapter 12
ICFJ
building an ethical AI use policy
How the International Center for Journalists turned staff disagreement about AI into a shared, written standard.
Chapter 13
SBP
intelligent ask amounts
GoFundMe Pro’s predictive ask amounts during hurricane response: +66% average one-time gift, +59% donor conversion.
Chapter 14
Suncoast Humane Society
small team, intelligent growth
A lean shelter raised nearly $37,000 in an off-season campaign, re-engaging donors lapsed five and six years.
Chapter 15