AI Readiness Copilot

Readiness Assessment

Sector-level AI readiness scores (0–100), normalized from public indicators. Each score carries a confidence level reflecting the recency and completeness of underlying data.

Sector readiness

AI Copilot

Ask policy questions in plain language. Answers are grounded in real data (World Bank, WHO, Mo Ibrahim IIAG) and powered by Mistral.

Answers are projections, not decisions. Human review required for final policy actions.

Try asking

Investment Simulator

Allocate a hypothetical budget across investment areas and see projected readiness gains as ranges — Conservative, Moderate, and Optimistic — rather than single-point predictions.

Projected sector readiness by scenario

SectorCurrentConservativeModerateOptimisticConfidence

Responsible AI Center

Transparency on data quality, what this system does not do, and when it defers to human review.

Data quality scorecard

SectorConfidenceMissing / flagged

What this system does NOT do

  • It does not allocate budgets or finalize investment decisions.
  • It does not produce single-point forecasts — only scenario ranges.
  • It does not account for local political priorities or community needs not reflected in the underlying data.
  • It does not generate a recommendation when confidence is below 60% or datasets conflict — it defers to human review instead.

Human-in-the-loop

  • A human policymaker or domain expert reviews, modifies, or rejects every recommendation before action is taken.
  • Recommendations include their reasoning and confidence level — never presented as certain.
  • When confidence < 60% or sources conflict, the system shows a "human expert review required" notice instead of a recommendation.

Review status —