Scores are judgment calls. If you have evidence that changes the picture — a missed event, a disputed classification, a suggested case — Sophie Krantz reads every submission.
A living index of organisations using influence to solve problems affecting billions — the model underneath each of them, and how to build it faster.
Hard problems persist in the world. Preventable blindness. Pandemic unpreparedness. Brain drain from entire regions. Three billion informal workers with no institutional infrastructure. The carbon markets that could finance forest protection, still too weak to function.
These problems share a structure: markets won't touch them because they aren't profitable, and states can't solve them because they require commitments that cross borders, decades, and political cycles.
Some organisations are solving them anyway — not by waiting for the incentive structures to change, but by building a different kind of power. Not financial leverage. Not political authority. Governance influence: the ability to change how other institutions behave through commitments so specific, so costly to abandon, and so evidently effective that others copy them without being asked.
This index was created by Sophie Krantz to make that model visible, measurable, and — critically — imitable.
Each actor profile shows three things: the scores across soft power, hard problem fit, and capacity; the founding decision that started the compounding clock; and the acceleration lever — the specific way that digital infrastructure, AI, or network effects could compress a decades-long timeline into years for anyone building this model now.
The aim is not to celebrate the organisations profiled. It is to document the pattern underneath them clearly enough that someone else can build on it — faster.
Scores are judgment calls, updated as events unfold. The index is openly maintained — events, actors, and disputes are submitted by contributors worldwide and curated editorially.
The index was sparked by Anthropic's decision in early 2025 to refuse a Pentagon request to relax AI safety guardrails — a moment that raised a larger question: which institutions are willing to spend their soft power on hard problems rather than narrow interests, and what does that model actually look like?
Organisational soft power: the ability of an institution to change how other organisations and governments behave, through governance choices that others voluntarily adopt. Not reputation. Not stated mission. Demonstrated influence on external behaviour — whether others have changed what they do because of how this organisation chose to govern itself.
Soft Power (SP) — 45%
Evidence of replication: standards embedded in other institutions' contracts, models built on in new geographies, approaches others have made their own. Highest weight because replication by choice is the clearest signal a model works — and the mechanism by which it scales.
Hard Problem Fit (HPF) — 35%
Specificity and difficulty of the named problem. Highest scores: problems ignored by markets and states, defined precisely enough for third-party verification, not solvable by commercial self-interest. The precision of the problem statement is the first indicator of seriousness.
Capacity to Deliver (Cap) — 20%
Track record, resources, institutional infrastructure. Weighted lowest — capacity without a hard problem and a real commitment produces nothing distinctive.
Index = (SP × 0.45) + (HPF × 0.35) + (Cap × 0.20)
Before scoring, each candidate is assessed against five criteria that determine index eligibility:
1. Legible refusal — Has the organisation declined something profitable, and can the specific decision be named?
2. Costly commitment — Is there a structural arrangement with observable consequences for violation — not a pledge or a target?
3. Hard problem specificity — Can the problem be stated in one sentence with a credible third party able to verify progress?
4. Model export — Has the model been copied by others who were not required to copy it?
5. Honest tension — Every profile names the tension that would lower the score if it worsened.
Every profile includes two additional fields that go beyond measurement: the founding decision — the specific early structural choice that started the compounding clock — and the acceleration lever — the concrete way that digital infrastructure, AI, or network effects could compress the same journey from decades to years for an organisation building this model now.
The index is designed to be used, not just read.