Turn real-world problems into progress

How a solution earns the word "solved"

Nobody can mark their own idea as validated. A solution only moves forward when there is a reason: a review, a real test, evidence, and a reviewer who checks that evidence.

1. Step 1
   Proposed

   Someone has put forward this idea. Nothing has been tested yet.

2. Step 2
   Under review

   The problem owner or team is reviewing whether this approach makes sense.

3. Step 3
   Ready for development

   The approach was accepted. It is ready to be built.

4. Step 4
   In development

   The team is building the solution.

5. Step 5
   Ready to test

   There is enough here to run a real test.

6. Step 6
   Testing

   A test is currently running.

7. Step 7
   Validation requested

   A test finished and the evidence is waiting for review.

8. Step 8
   Validated

   A reviewer checked the evidence and confirmed the solution works.

9. Step 9
   Impact

   The validated solution is producing measurable real-world results.

Trust & verification

AI researches. Humans decide. ProblemBank coordinates.

Every problem, solution and impact claim follows the same master lifecycle. AI research gathers evidence and context, but it never approves, rejects or verifies anything — that decision is always made by a ProblemBank reviewer. Contributors and teams do the building and testing, but nobody can validate their own solution, mark their own work solved, or confirm their own impact. Only impact that ProblemBank has verified is counted publicly; everything else is shown as "Reported — awaiting ProblemBank verification".

• AI researches

  Gathers evidence, checks for duplicates, and drafts a report — never a decision.

• ProblemBank verifies

  A human reviewer reads the AI research and evidence, then decides what happens next.

• Contributors build and test

  Teams develop and test solutions, but every verification step is done by someone else.