Children's Healthcare of Atlanta Foundation, Children's Hospital of Pittsburgh Foundation, and St. Jude Children's Research Hospital/ALSAC are already using AI in fundraising, but in different ways: one foundation analyzed five years of clinical-family data to surface high-capacity patient families, another used the same vendor for donor identification, and ALSAC is using Copilot and ChatGPT to speed donor research and stewardship writing. A Kindsight/Cherian Koshy survey of 510 healthcare donors adds the donor-side check: 69.6% said AI can improve charitable effectiveness, 21.5% would increase giving if an organization uses AI, and 87% said transparency about AI use is crucial for trust [1][2][3][4].
Two AI paths are already in use
The tools in play separate cleanly into two jobs. Predictive AI tries to rank people: who may have giving capacity, which patient families deserve a closer look, and which names should move up the prospect list. Generative AI tries to compress the work around those people: research notes, outreach drafts, stewardship messages, and other repetitive writing that slows gift officers down.
- Predictive AI: donor propensity scoring, grateful-patient identification, and prospect prioritization.
- Generative AI: donor research, first-draft communications, and stewardship personalization.
Predictive AI is the deeper operational bet
Children's Healthcare of Atlanta Foundation is the clearest public example. DonorSearch says it analyzed five years of clinical-family data, surfaced previously undetected patient families with $5M+ giving capacity, and helped create a new Office of Transformational Giving [1]. That is the kind of evidence that matters because it shows AI moving from a score into a workflow, not just into a dashboard.
Children's Hospital of Pittsburgh Foundation used DonorSearch for data-driven donor identification as well [2]. The public detail is thinner than the Atlanta example, but the operational point is similar: predictive scoring is moving into the daily work of prospecting, where gift officers decide whom to call, visit, and steward next.
That is also where the ethical discomfort starts. The closer the model gets to patient-adjacent information, the more important it becomes to know what the score means, who can see it, and what happens when it is wrong.
Generative AI is mostly a workflow accelerator
St. Jude Children's Research Hospital and ALSAC offer a different deployment. According to CloudWars, the organization uses Copilot and ChatGPT to speed donor research and personalize stewardship communications, which leaves gift officers more time for relationship-building [3].
That use case is narrower than predictive scoring, but it is easier to defend. It does not decide who belongs in the pipeline; it shortens the time between a relationship and a response.
The promise is bigger than the proof
DonorSearch's client data claims up to an 85% increase in response rate, 81% accuracy for repeat-donor prediction, 20% higher average gift amounts, and a 17% faster major-donor pipeline [5]. Those numbers are attractive, but they are vendor-reported rather than independently validated, so they should be treated as a sales signal, not as a field-wide benchmark.
The maturity gap shows up in the wider health philanthropy sample too. In the 2026 Jarrard Pulse Check of 33 healthcare philanthropy professionals, most teams rated their AI competency as moderate and said they use AI more for communications than for strategic prospect identification [6]. The 2026 CCS Philanthropy Pulse, based on 52 health organizations, reported 63% revenue growth but donor retention falling from 40% to 34%, which is a reminder that growth and durability are not the same thing [7].
Legitimacy comes from transparency
The donor survey points in the same direction: people are not uniformly opposed to AI, but they want to know when it is being used. That matters more in children's hospitals, where family and patient-adjacent data can feel intimate even when it is technically permissible. The practical floor is straightforward: HIPAA-aware architecture, an internal AI policy, and plain-language donor disclosure where AI changes research, scoring, or messaging. AI for children's hospital fundraising looks most credible when it helps locate overlooked prospects or reduce stewardship drag, not when it is sold as a replacement for judgment. The strongest cases show real operational promise, but the field still needs independent validation and visible transparency before children's hospitals can treat AI as a durable fundraising advantage.
References
- DonorSearch case study: Children's Healthcare of Atlanta Foundation
- DonorSearch case study: Children's Hospital of Pittsburgh Foundation
- CloudWars article: St. Jude Children's Research Hospital / ALSAC
- Kindsight AI and Donor Perceptions Report
- DonorSearch client data
- 2026 Jarrard Pulse Check
- 2026 CCS Philanthropy Pulse
Comments
Join the discussion with an anonymous comment.