Tout le monde utilise l’IA; presque personne n’en est transformé

In February 2026, Virtuous and Fundraising.AI published a survey of 346 nonprofits that deserves careful attention. Its central finding is paradoxical: 92 per cent of organisations now use artificial intelligence in some capacity, yet only 7 per cent report a major improvement in their organisational capability. Canadian data point the same way; Blackbaud’s Status of Canadian Fundraising 2026 reports that roughly nine in ten charities use AI, while sophisticated applications remain uncommon.

The efficiency plateau

The survey’s authors describe an “efficiency plateau.” Staff use AI to draft emails and summarise documents faster, which is genuinely useful, but the organisation itself works much as it did before. Eighty-one per cent use AI individually and on an ad hoc basis; only four per cent have documented, repeatable workflows; and nearly half have no formal governance policy at all.

This pattern is familiar to anyone who has led technology change. A tool adopted by individuals improves individual productivity. A tool adopted by an organisation — with shared processes, clear ownership and measured outcomes — changes what the organisation is able to do.

From personal convenience to institutional capability

Moving from the first state to the second requires neither large budgets nor technical brilliance. It requires deliberate choices:

• Name the problem before the tool. Begin with a bottleneck — grant reporting, volunteer scheduling, donor acknowledgements — and ask whether AI can relieve it.
• Write the workflow down. A repeatable process, documented in a page, is worth more than a dozen clever prompts known to one person.
• Govern early. Decide what data may and may not be entered into AI tools, who reviews outputs, and how errors are corrected.
• Measure something. Hours saved, turnaround time, response rates: without a measure, improvement is merely an impression.

A note of encouragement

Interestingly, the report found that smaller organisations were more likely than larger ones to report moderate impact. Agility, it seems, can compensate for scale. For the many small charities we serve, that is welcome news: the path from experimentation to genuine capability is shorter than it appears, provided one walks it with intention.