What the publication is
An independent evaluation of the African Development Bank’s Selectivity Framework, covering 2013 to 2024 and completed in April 2026. The evaluation draws on a portfolio of more than 2,500 operations, seven country case studies, key informant interviews, and a benchmarking study against other multilateral development banks.
Its inception report, peer reviewed inside IDEV and by the European Bank for Reconstruction and Development, sets out an AI-assisted methodology: evidence from four background reports is coded and extracted against the evaluation questions, with Evidano named as the tool to validate the findings and every AI output subject to manual review and triangulation.
How Evidano fit the workflow
- 1Collect the evidence base
Development literature, strategy documents, reports, and interview notes.
- 2AI assisted coding and extraction
Evidence is coded, extracted, and summarized, mapped against the evaluation questions in a structured database.
- 3Human review and triangulation
Outputs are checked manually and cross-checked with traditional methods and expert analysis.
- 4Findings validated
Evidano is named as the tool to validate findings, trained for development literature.
“...an AI-powered evaluation tool... specifically trained to analyze and evaluate development literature.”
“Manual review of AI outputs will be conducted to validate findings.”
“AI will be used as a complementary tool, never a replacement for expert human judgment.”
AI insights are triangulated with traditional methods and expert analysis, with manual review of AI outputs before findings are validated.
Evidence is coded and mapped against the evaluation questions in a structured way, helping keep analysis consistent across sources.
The evaluation states AI tools operate offline or in house, data is anonymized, and informed consent is obtained before recording interviews.
The inception report was reviewed by IDEV evaluation officers and an external reviewer from the EBRD evaluation department.