How public services are learning to use their data.

The trope of the “data-driven public sector” is older than most of the software it refers to. What has changed, in the last decade, is that a handful of institutions have stopped treating data as a reporting layer and started treating it as an operating input. Five patterns worth paying attention to.

Five institutions, five patterns

1. 01France - L'Assurance Maladie replaced random audit sampling with pattern-based detection. The system flags claims whose shape is atypical, and investigators spend their time where it actually matters.
2. 02United Kingdom - NHS England runs a predictive layer over consultation data to anticipate regional flu outbreaks. Medication stocks and staffing are reallocated before the wave arrives, not after.
3. 03Finland - Kela, the Finnish social insurance institute, looks across a citizen's medical history to identify patterns associated with long-term work disability. When the pattern appears, preventive follow-up is triggered before the situation deteriorates.
4. 04Singapore - The Central Provident Fund uses a mobile app to proactively notify citizens of social benefits they are entitled to but have not claimed. Entitlements that used to be missed through paperwork are now pushed to the user.
5. 05Estonia - The pension agency publishes a personal retirement simulator pre-filled with each citizen's career data. Choices - contribution level, retirement age, career breaks - update the projected pension in real time.

What these five have in common

None of these are technology projects dressed up as public policy. They are the other way around: policy decisions that happen to be implementable because the data is already there. The institutions that will close the gap on public trust in the next decade are the ones that take that distinction seriously.