A charity gets named by ChatGPT as a trustworthy place to give for four specific reasons — its evaluator ratings, its documented effectiveness, its earned-media coverage, and the accuracy of its source-layer presence. The forthcoming Nonprofit AI Citation Share Study isolates those four drivers across more than 600 answer-engine responses. Every one of them is buildable. Generative Engine Optimization — GEO — is the discipline of building them on purpose.

GEO is what search optimization was for the era of ten blue links: the practice of making an organization the answer rather than a result. For a nonprofit, the answer that matters is the response to "best charity for [cause]" — and winning it is mechanical, not magical.

The four drivers — and the proof

Evaluator ratings. A current, strong Charity Navigator rating, a GiveWell recommendation, a Candid transparency seal, a BBB accreditation. This is the strongest single driver. The study found the engines route the trust question through the evaluators — Charity Navigator alone captures 38 percent of evaluator citations. The rating is the input; the citation is the output.

Documented effectiveness. Against Malaria Foundation, GiveDirectly, and Direct Relief over-index in the study — cited far above their budget size — because their impact is documented in a form the engines retrieve. In the global-health cause, GiveWell-associated charities capture the entire top of the table. Documented effectiveness beats name recognition.

Earned-media coverage. Sustained, accurate press coverage feeds the source layer. World Central Kitchen's strong disaster-relief citation share is substantially an earned-media effect.

Source-layer and entity coverage. Charities with complete, accurate, well-structured presence — across evaluator profiles, their own site, and the structured entity data the engines read — are retrieved confidently. Thinly documented charities are simply absent.

The 90-day playbook

GEO for a nonprofit is a sequence, not a campaign.

Audit and complete every evaluator profile. Charity Navigator, Candid, GiveWell where applicable, BBB. Current financials, complete narrative, every available data field filled. This is the highest-leverage move and most organizations treat it as an afterthought.

Document the impact. Outcome data, independent evaluation, cost-per-result where it exists — published in clear, retrievable form on owned properties, not buried in an annual PDF.

Saturate the source layer. Earned coverage, accurate presence in comparison and cause-roundup content, a complete structured entity record so the engines retrieve a consistent account.

Audit and re-run. Run the trust-layer queries across the five engines, measure Citation Share, find the gaps, fix them, measure again. Full methodology: The Six-Step GEO Citation Audit.

What not to do

Do not treat GEO as a slogan exercise. The engines do not retrieve a tagline or a mission statement. They retrieve evidence — ratings, outcomes, coverage. A charity that calls itself "the most trusted" without the evaluator standing and documented impact behind it moves nothing. The work is the evidence; the positioning follows from it.

The GEO framework: The GEO Canon · GEO in One Page · Six-Step Methodology

Sibling sector playbooks: GEO for Churches · GEO for Healthcare · GEO for Content Teams

Nonprofit coverage: The Nonprofit AI Citation Share Study