Who it’s for · Funders · Private philanthropy
Know how well your giving is working, without turning it into a bureaucracy.
You give because you want lives to change. Underneath the giving is a question worth answering honestly: how much of it is working. Prova is being built to give you an independent read, held in confidence, that respects the people you fund and never turns your philanthropy into an institution.
The bind
Renewal incentives can make an independent answer harder to obtain.
People seeking renewed support face an incentive to emphasize success. Advisors and donors can also hesitate to press a question that feels like distrust of people they have chosen to believe in. These pressures do not make the reports false. They make an independent read useful when the decision matters.
Bespoke impact evaluation can be difficult to justify for an individual grant. The result is often a decision made from relationships, output reporting, and partial evidence.
What Prova does
An honest read, held close, that respects the people you fund.
Prova Method gives you an independent answer to a simple question: how much is your giving actually changing. It starts from what the organizations you fund already record. Additional work depends on the quality, coverage, permissions, and form of those records. It establishes the strongest claim the evidence honestly supports, and tells you plainly where it falls short.
Prova Method does not run the programs it assesses, and its fee does not depend on the conclusion. What it finds is yours, held in confidence subject to the agreed legal and safeguarding boundaries, for your judgment, and it goes no further without your word.
None of this asks you to build a department or become a foundation in the bureaucratic sense. It is meant to give you a stronger basis for deciding what to continue, change, or examine further. AI does the labor-intensive reading and synthesis behind it. Prova is being built to establish how far that capability lowers the cost of the full work. The judgment, and the care owed to the people in the records, stays human.
How to start
Start with one program.
A bounded first step can assess what one program’s existing evidence supports, with no commitment beyond it. Timing and additional work depend on the question, records, permissions, and design. Any continuing measurement function is separately scoped after that first engagement.
Start a conversation
Start with a conversation, in confidence.
If you would rather know than wonder, that is the conversation to have. The first one is a fit check, with no obligation, and I take these myself.
Common questions
The questions a careful reader asks first.
Is cheaper evidence weaker evidence?
A lower price does not require a lower standard. Frontier AI models can now perform much of the reading, extraction, and synthesis in bounded tasks. Prova is being built to establish how far that capability lowers the total cost of reaching a given grade in its claim-evidence rubric without shifting work or risk elsewhere. Design labels do not guarantee credibility. A modest claim remains modest, and a thin record remains thin.
AI can be confidently wrong. Why trust it here?
It can, and that is a central design problem. AI does the labor-intensive reading and synthesis; people decide whether a finding holds, what it means here, and where the evidence runs out. Prova is being built so claims retain provenance and uncertainty. Its reliability across real engagements is still being tested.