Let's talk about Finland's structures.
AI reviews laws and grants. Surfaces the ones worth discussing.
Legislation Analysis
Every Finnish law gets a verdict: keep, simplify, or repeal. AI reads the law text and assesses its impact on economic growth.
View law analysis →STEA Grant Analysis
€274M distributed to 686 organizations. We reviewed every org and every grant. Finding: €44-65M could be saved.
View STEA analysis →Method
Analysis based on public data (Finlex, STEA) and Claude Opus 4.6 AI model. Everything analyzed by Opus: law texts, grants, overlaps and savings calculations.
Law analysis
- Data: Finlex Open Data API
- Every law text read by Claude Opus 4.6
- Verdict: keep, simplify or repeal
- Cost estimates are order-of-magnitude
STEA analysis
- Data: avustukset.stea.fi + RAY 2000-2016
- 766 applications, 1527 grant lines analyzed
- 51 overlap groups identified
- 27 years of funding history
✓ Law texts and grant amounts are real
~ Assessments and classifications are calculated
FAQ
Organizations need administration to function. Why is that a problem?
Administration isn't the problem. 686 separate administrations when 100 would suffice is the problem. Each org needs its own board, audit, reporting. Merging would save admin costs without reducing services.
Regional presence matters. Wouldn't merging weaken it?
Regional work would continue as local offices - like wellbeing regions merged municipal health services. Staff and services stay, admin gets lighter.
General grants give flexibility. Isn't that good?
Flexibility is good, but 21% of total funding with zero performance targets is too much. A reasonable level would be 20-30% per org, not 60-100%.
Is this a political project?
No. This is data analysis from public sources. We don't represent any party or interest group. Data is open and anyone can verify.
Who made this?
Mari Luukkainen (@mariluukkainen) and Claude Opus 4.6. Code is on GitHub.