AI maturity in Swiss nonprofits: what the research shows — and what it doesn't

There is no reliable figure: no published, quantitative study measures AI maturity in the Swiss nonprofit sector. Anyone wanting to know how widespread AI is in Swiss nonprofits has to triangulate from three neighbouring fields — the Swiss economy as a whole, the nonprofit studies from Germany and Austria, and international surveys. And that triangulation reveals a remarkably consistent pattern: high individual use alongside an almost universal absence of strategy. AI is already being used, but it is barely being led. This very vacuum — not the use — is the real finding.

The key points at a glance

• Active use is growing fast — in the Swiss economy. Active AI integration in Swiss SMEs jumped from 22% (2024) to 34% (2025) (AXA/Sotomo SME Labour Market Study 2025). That is the economy as a whole, not the nonprofit sector.
• Individual use is high, organisational embedding is shallow. 80% of Swiss professionals and managers have personal experience of generative AI, yet only 23% report an organisation-wide use in data analysis (ZHAW AI Barometer 2025).
• The best nonprofit figure comes from Germany. 73% of volunteers and staff in the charitable sector use generative AI — with only minimal strategic embedding (ZiviZ/Stifterverband 2025, Germany, not Switzerland).
• The strategy deficit is measurable. In Austrian nonprofits, 32% deploy AI tools, but 78% have no AI strategy (npoAustria/WU Vienna 2024, Austria, not Switzerland).

What the Swiss figures tell us

For the Swiss economy as a whole, the data are good — they just say nothing about nonprofits. The AXA/Sotomo SME Labour Market Study 2025 shows how quickly the Swiss economy is adopting: the share of SMEs with active, deliberate AI integration rose within a single year from 22% (2024) to 34% (2025), while the share forgoing AI fell from 45% to 29%. That is the pace against which any nonprofit lag could even begin to be measured.

More revealing still is the ZHAW AI Barometer 2025, because it lays bare the real fault line. 80% of the Swiss professionals and managers surveyed have personal experience of generative AI, and 93% use AI to generate text. But the potential is unlocked individually, not organisationally: 96% see potential in data analysis, yet only 23% report a genuinely organisation-wide use. It is a small survey (n=137), so more an indication than a representative rate — but it evidences, for Switzerland, precisely the diagnosis suspected in the nonprofit sector: high use, shallow embedding.

The oversight perspective comes from the swissVR Monitor 2024 (Deloitte, HSLU, n=391 Swiss board members). 70% of boards have engaged with generative AI, and the greatest perceived risk is flawed AI results (60%). Yet only 17% have all AI outputs reviewed by a human, and roughly three in four receive scarcely any or no regular reporting on AI use. These figures, too, apply to companies in general — not to nonprofits. What transfers is the structure of the problem, not the measured value.

What the nonprofit studies from the DACH region show

The moment the focus turns nonprofit-specific, you leave Switzerland behind. The best figure currently available in the German-speaking world comes from ZiviZ im Stifterverband (2025) — an online survey of n=1,235 volunteers and staff in the German charitable sector, conducted with the Swiss VMI of the University of Fribourg as partner. The result: 73% use generative AI, above all for text work, research, translation and communication — but only a minority of organisations have guidelines, strategies or clear responsibilities. The study's title is at once its finding: "individually used, but strategically neglected".

The Austrian study by npoAustria and WU Vienna (2024) measures the same thing at the organisational level: 32% of nonprofits already deploy AI tools (again, above all for text creation, research and translation), but 78% have no AI strategy and no planning. 52% expect strong to very strong change from AI and digitalisation — and the greatest need cited is building staff competence.

Both studies are explicitly not Switzerland. Whether their figures transfer is an open question — and there are good reasons for caution. The Swiss militia system shapes nonprofit boards differently from the professional, staffed structures in Germany; the revised Data Protection Act (revDSG) sets its own framework; and multilingualism changes how, and with which tools, work is done at all. The DACH figures are the best available approximation — and no more than an approximation.

The pattern across all sources

What stands out is the robustness of the finding across wholly independent surveys. The Swiss economy as a whole (ZHAW), the German nonprofit sector (ZiviZ) and Austrian nonprofits (npoAustria) come from different countries, with different methods, to the same diagnosis: use is high, governance does not follow. Internationally, the Salesforce Nonprofit Trends Report (7th edition, 2025) confirms the picture — more than half of the nonprofits surveyed worldwide are piloting or actively using AI, while governance lags behind.

When a pattern says the same thing in four independent contexts, it is more dependable than any single figure. For Swiss nonprofits that means: you cannot cite a Swiss percentage, but you can say with high confidence what kind of problem is at hand. This is not an adoption problem. It is a leadership problem.

The research gap: what no study answers

Clear as the pattern is, for the Swiss nonprofit sector four questions remain concretely unanswered:

1. No Swiss nonprofit AI maturity survey. No published study measures AI adoption or governance representatively among Swiss associations, foundations and federations. Every nonprofit-specific figure comes from Germany or Austria.
2. No maturity segmentation. Even the DACH studies measure "use yes/no", not a multi-stage maturity level — from ad hoc through piloted and strategically embedded to scaled. A nonprofit-suited maturity model built on Swiss data does not exist.
3. A governance blind spot in the nonprofit context. The swissVR Monitor supplies oversight figures for companies only. For volunteer foundation boards and association boards under the militia principle, there are no Swiss data on AI oversight, policies or competence.
4. No differentiation by size and resources. ZiviZ shows strong differences by organisational size — for the fine-grained Swiss nonprofit landscape, with its many micro-associations, the corresponding data are entirely absent.

For a nonprofit board, this is no academic footnote. It means that decisions about AI are being taken today without a reliable evidential basis — neither about one's own position nor about the comparison with similar organisations. Anyone waiting for "the study" that finally puts a number on Swiss nonprofit maturity is waiting for something that does not exist.

What Swiss nonprofits should take from this

A data gap does not entail an action gap. Four pragmatic conclusions can already be drawn today:

1. Assess your own position rather than wait for studies. The relevant figure is not the sector's but your own organisation's. An honest internal baseline assessment — who uses what, for what, with what safeguards — yields more than any industry average.
2. Take stock of shadow AI. If 73% of volunteers in the DACH region already use AI, your staff most probably do too — often without the leadership knowing. A simple inventory of all AI tools in use, including the AI features in bought-in software, is the first honest step.
3. Governance before scaling. The consistent pattern — high use, missing strategy — is an invitation to do it better: guardrails first, then expansion. Which questions a board must ask for this is covered in our article on AI strategy in the boardroom.
4. Build competence deliberately. The greatest need cited in the studies is building competence — and for organisations with an EU nexus, AI literacy is already a legal obligation (see EU AI Act for Swiss nonprofits). For militia structures, a pragmatic, lightly documented training of the active members often suffices.

The research, then, does not tell you where your organisation stands — but it tells you very precisely which question you should be asking. Not "are we using AI yet?", but "are we leading it too?".

The first step is an honest baseline assessment. If you would like to know where your organisation truly stands today, arrange a free initial consultation — or take a look at the path to effective AI, from baseline assessment to secure operation.

This synthesis is based on publicly available studies; last reviewed July 2026.

Harry Witzthum

Founder of Rautaki · Doctor of Philosophy · NPO manager VMI