Every price on this page is a published list price with a source. We are neither the least expensive option nor the largest. The arithmetic follows, so you can decide without a sales call.
Three findings define the gap this practice sits in. None of them is about enthusiasm for AI.
The Center for Effective Philanthropy surveyed 215 foundation leaders and 451 nonprofit leaders in April 2025. Only about 10% of foundations report supporting grantees on AI implementation, and almost two-thirds of foundations and nonprofits say none or just a few of their staff have a solid understanding of AI. source
Published list prices as of September 2026, in each provider's own currency, sourced underneath. Where a provider publishes no price, the row says so, which is itself a finding.
Excludes no one, which is why it prevails by default.
Excludes no one. The constraint is capacity, not eligibility.
US-based nonprofits only. Not open to organisations in Europe, MENA or the international sector.
Under 100 staff and under USD 10M budget, in named sectors. Excludes larger INGOs, UN bodies and public institutions.
Excludes no one, though it certifies one person rather than the organisation.
Excludes no one, and that is the difficulty. Nothing in it knows your data rules.
Excludes no one. The same catalogue for every organisation.
Not built for beneficiary data, donor disclosure or safeguarding vocabulary.
Reaches your director. Does not reach the officer who pastes the file.
Qualifies a person to write a policy. Does not write one.
Governance-first. Trains no one, so behaviour does not change.
Will not take a CHF 20,000 engagement. You are not their client.
Organisations needing model building, data pipelines or thirty trainers next month.
Sources, checked September 2026: NTEN certificate · NTEN cohort · NetHope Accelerator · LinkedIn Learning · Coursera for Teams · Section · executive programmes · IAPP AIGP · governance consulting ranges · CEP, AI With Purpose. Providers change prices; verify before deciding. Volume discounts are common on per-seat platforms and are not reflected in list prices.
These are line items you can find in your own accounts and your own grant calendar.
while few are helping grantees prepare: only about 10% of foundations report supporting grantees on AI implementation, and almost two-thirds of foundations and nonprofits say none or just a few of their staff have a solid understanding of AI source
The arithmetic here concerns revenue rather than productivity. One mid-size foundation grant is worth more than this entire engagement, usually by a multiple. If an unanswerable AI-governance question delays a single application by one cycle, the cost of that delay exceeds the cost of never having the gap. The policy, the tool inventory and the assessment report are the answer to that question, and they take eight weeks to produce, which is shorter than most grant cycles and longer than most deadlines allow once the question has already arrived.
A full-time grant writer runs roughly USD 98,000 a year fully loaded. Outsourced, a foundation proposal is USD 1,500–5,000 and a federal application USD 5,000–15,000+. Most organisations pay neither. They absorb it as programme staff time, which is why it never appears as a line and never gets managed.
Why this is measurable and most AI claims are not. You already know how many proposals and donor reports you produced last year. Count the hours on the next three before we start and on three after. That is the number, and it comes from your own accounts. We will not put a percentage in a proposal, because an ROI figure for AI training quoted before your documents have been seen has no basis.
Where AI use is individual and undocumented, the question has no answer inside the organisation. In a benchmark of 346 nonprofits, 81% use AI individually, without shared workflows, and 47% have no AI governance policy at all. source If nobody has asked, nobody knows, and the exposure is carried without sight of it.
For you this is not a fine. It is a safeguarding failure involving the people you exist to protect, discovered by a donor or a journalist rather than by you. There is no number on that. There is a number on preventing it, and it is on this page.
None of these is “we are more tailored.” We have sorted them by how hard they are to copy, and we say so, because a buyer who sees that admission can trust the rest.
The best free programme is open to US-based nonprofits only. The main subsidised accelerator caps at under 100 staff and under USD 10M in named sectors. A Geneva international body, a 200-person INGO, a public institution or a MENA organisation is outside all of it. We are not competing with free, because free is not available to you.
Ten dimensions, sixty scored questions, mapped to ISO/IEC 42001, the NIST AI RMF, the EU AI Act, GDPR, OECD and UNESCO instruments. Published free. Every organisation scored adds to a dataset that, over time, becomes sector benchmarking no one else holds.
Donor reporting, beneficiary data, safeguarding, disclosure to funders, DPIA, and delivery in English, French or Arabic without translation. Certified specifically in teaching AI fluency and in AI fluency for nonprofits, as well as in using it. Others can get the certificates. The combination, based in Geneva, is rare.
A lapsed licence leaves nothing. This leaves a policy in force, a page per department, a tool inventory, an attendance record and two scores. Others could work this way; platform business models structurally cannot.
Platforms train but write no policy. Governance consultancies write policy but train no one. A governance firm could add training tomorrow. In this price band, none has.
Anyone could show their prices before a call and charge once per organisation instead of per person per year. Almost no one does: platforms because it breaks their model, consultancies because it removes their negotiating room. We do it because not knowing whether this is a CHF 5,000 or a CHF 500,000 conversation is the main reason buyers leave.
If you need model fine-tuning, data engineering or custom software, we do none of it. We configure agents and tools on the platforms you already use, and stop there. If you need thirty trainers across four countries next month, we cannot staff it. If your team has fewer than ten people, a small-group course fits better than a full engagement, and we will say so on the call. Those limits are real.
The assessment scores you across ten dimensions and tells you what to fix first. You keep the report whether or not you ever speak to us.