Insights from 16 nonprofits in the OpenAI People-First AI Fund, and how much they turned out to have in common.
Read the paperSix nonprofits running five months of agentic fundraising in the open, with everything they learn published as they go.
Meet them
25 Aug 2026No supporter gets the same email twiceWhat personalized email actually runs on, what we capture, and what we're testing.
18 Aug 2026Your donor used AI to give to your causeA new Senate bill, a payments protocol, and what agentic giving means for consent.
12 Aug 2026Six nonprofits selected to pioneer agentic givingFive months of open experiments in agentic fundraising, published as we go.AI personalized
No supporter gets the same email twice.
Outreach runs on an evidence library of peer-reviewed donor-behaviour research rather than copywriter instinct. The decision engine reads each supporter's profile, draws the findings that actually apply to them, and builds the journey — inbox, campaigns, forms, events and the send calendar all working from the same record.
More than 200 peer-reviewed and tested studies of donor behaviour sit behind every sequence, with around ten drawn for any one campaign. A finding that does not replicate when the cohort tests it comes back out of the library.

Most files carry years of dormant contacts. Enrichment and cleaning separate the genuinely engaged from the never-asked — across the cohort that meant working down from more than 200,000 contacts — and re-engagement opens with a plainly written note and an optional two-minute quiz rather than another ask.

"Generate me a campaign for the people who joined our Climate Week event." The engine reads each supporter's profile and writes the sequence around them — one campaign, different content per person, every message tracing back to the profile and the findings that produced it.

A thin record is filled rather than guessed at. Public context is added only with explicit consent, and where that still is not enough the platform goes and asks — the way Net Impact is building impact-themed quizzes to re-engage alumni now established in their careers, turning a two-minute answer into the brief that writes their next email.
Automated AI journeys start by asking. As a routine proves itself it earns the right to run on its own, built from the logic and evidence of real donor interactions and your playbooks — while higher-value decisions still come to you for review. How often people override it is measured, not hidden.

Whitelabel vs the market
A quick, honest look at how Whitelabel stacks up against ChatGPT plus a CRM and Mailchimp for nonprofit teams.
WhitelabelComparison shown for illustration. Competitor terms change, so always check current pricing.