Trust & your data

We never receive your donor records.

We do not train models on your materials. And every claim in a campaign traces to a source you gave us — nothing invented.

What we receive

Brand and voice documents you share (appeals, newsletters, case statements)
Audience segment labels and counts (e.g. “lapsed young alumni”)

What we never receive

Donor or alumni names
Email addresses and contact records
Giving history and gift amounts
Any individual donor record

How your documents are used

The documents you share are used only to build your voice profile and the provenance corpus for your campaigns. They are not used to train foundation models, and they are not shared across institutions.

Provenance, not hallucination

Every factual claim in a campaign is traceable to an approved source in your own document corpus. The AI does not invent facts about your institution, and nothing sends without a human on your team approving it first.

See it in the sample campaign →

Security posture

Data is hosted with reputable cloud providers, access is restricted to what each task needs, and a human approval gate stands before anything is exported for sending. For procurement questions — including higher-ed reviews (HECVAT) and our position on FERPA and SOC 2 — reach us at hello@givgiv.org.