Supporting case
Qonto’s transaction data show growth in paid AI subscriptions, with clear limits on what that measures
What happened
Qonto’s October 21 explanation reports that the share of its business customers paying for at least one AI tool rose from 9.5% in September 2025 to 16.7% in September 2026. That is substantial growth, but it is not a literal doubling. The underlying index and methodology ↗ draw on anonymized transactions across more than 750,000 Qonto business customers in France, Germany, Italy and Spain, including freelancers and smaller firms.
A paid subscription recorded through Qonto is a narrow observable behavior. It excludes free use and may miss embedded services or payments made elsewhere. The customer base is not a random sample of every European firm, and purchasing a subscription does not establish productivity gains, effective integration or independent implementation knowledge.
Paid AI subscriptions are a bounded uptake indicator. They do not demonstrate effective implementation, productivity or independence from vendors.
Dates and implementation stage
Development/status: 2026-09; source publication 2026-10-21
