Generative Ai Methodology for Uncovering Entrepreneurial Journeys in Africa

Entrepreneurship is often named as a potential driver of economic growth, innovation, and job creation in Africa. However, data on entrepreneurship in Africa remains fragmented. This study addresses that gap by leveraging generative artificial intelligence (GenAI) and a retrieval-augmented generation (RAG) framework to systematically analyze unstructured data sources (YouTube videos, podcasts, and reports) as they relate to startup founders. Specifically, this paper shows how those tools can be used to analyze entrepreneurial journeys by generating narratives of billion-dollar startup (i.e., unicorn) founders in Africa. In the process, our software solution identifies pivotal markers such as funding milestones, mentorship access, and the ecosystem interactions that shape entrepreneurial journeys. Our RAG methodology incorporates multi-query expansion, context retrieval, and structured narrative generation, with evaluation metrics focusing on retrieval accuracy (precision, recall, F1 score), generation quality (faithfulness, hallucination rate), and narrative completeness. Results demonstrate that high-quality founder narratives can be constructed even from limited but thematically diverse data sources. Methodologically, this study demonstrates the utility of GenAI in enabling large-scale qualitative research, providing a scalable approach to analyzing unstructured, multimodal data. The methodology overcomes challenges such as data fragmentation, transcription inaccuracies, and multi-speaker content attribution.