A nonprofit's visibility inside ChatGPT, Claude, Perplexity, Gemini, and Google's AI Overview is decided, to a degree most organizations have not registered, by Charity Navigator and GiveWell. The Nonprofit Citation Share Study 2026 found the answer engines route the donor's trust question through the evaluator layer first — Charity Navigator captures 38 percent of evaluator citations, GiveWell 24 percent — and then name the charities those evaluators rate well. The evaluator profile is no longer a back-office compliance file. It is a primary communications asset.

If a development team manages one thing for AI visibility, it should be this.

The evaluator layer, by the numbers

The study measured citation share among the authority sources the engines route through:

• Charity Navigator — 38 percent. The default evaluator the engines cite for whether a charity is trustworthy and efficient. It rates more than 230,000 nonprofits on a star and 0-100 system; the breadth and simplicity make it the most retrievable authority in the segment.
• GiveWell — 24 percent. The evaluator the engines cite for effectiveness and impact-per-dollar — the answer to "most effective charity," not "most efficient." Its trust-layer influence runs far ahead of the number of organizations it covers.
• Candid (GuideStar) — 16 percent. Profiles, Form 990 data, transparency seals.
• BBB Wise Giving Alliance — 13 percent. Twenty-standard accreditation.
• CharityWatch — 9 percent. Independent watchdog grades.

Across every one of the five engines, the route to a charity recommendation runs through these organizations.

Why this makes the evaluator profile a communications asset

A charity's Charity Navigator rating is built largely from its IRS Form 990 and the data it submits through the evaluator's portal. Its GiveWell standing depends on documented, evidence-grade impact. Its Candid profile depends on what the organization uploads and verifies. Each of these is partly within the nonprofit's control — and each is an input the answer engines retrieve.

That reframes the work. An evaluator profile that is incomplete, out of date, or thinly populated is not a minor administrative gap. It is a direct drag on the organization's visibility in the single place donors now begin their research. A complete, current, accurate profile is the opposite — a retrievable trust signal the engines surface.

What to do

Audit every evaluator profile. Charity Navigator, Candid, BBB, and GiveWell where the cause and evidence base fit. Find what is missing, stale, or thin.

Complete them. Current financials, full narrative, every available data field, the portal submissions each evaluator offers. A complete profile rates better and retrieves better.

Keep them consistent. The facts in the evaluator profiles, on the organization's own site, and in its structured entity data should all agree. Inconsistency weakens retrieval.

Make it joint-owned. Evaluator standing has historically sat with finance or development. In the answer-engine era it is also a communications asset, and the communications and development functions should own it together.

Common questions

Does my Charity Navigator rating really affect AI search?

Yes. The study found the engines route the donor trust question through the evaluators before naming charities. A stronger, more complete evaluator profile is a direct input to a nonprofit's answer-engine visibility.

Which evaluators matter most?

Charity Navigator and GiveWell capture the majority of evaluator citations — Charity Navigator for efficiency and trustworthiness, GiveWell for effectiveness. Candid and the BBB matter as well.

Who should own the evaluator profiles?

Communications and development jointly. The data is financial and programmatic; the visibility consequence is a communications outcome.