Case Study: World Bank Independent Evaluation Group achieves faster, scalable portfolio evaluation insights with Octant AI

World Bank Independent Evaluation Group - Customer Case Study

The World Bank Independent Evaluation Group engaged Octant AI in 2020 to explore whether artificial intelligence could improve the speed, scale, and efficiency of evaluating a large portfolio of 392 nutrition projects across 64 countries. The team faced a difficult, time-intensive process of reviewing lengthy completion reports, extracting both numerical and text-based evidence, and synthesizing lessons learned from a portfolio representing more than USD 28.8 billion in commitments.

Octant AI used its Octant AI Prevent product, along with data analytics, natural language processing, unsupervised machine learning, and expert-guided content labelling, to analyze nearly 1,000 reports and uncover patterns in project success and failure factors. The pilot showed strong results, including 90–95% accuracy at high-level content labeling and the identification of nine unique data points from five countries across more than 14,000 data points, demonstrating that Octant AI could help the World Bank improve evaluation speed, insight, and future project planning.

World Bank Independent Evaluation Group