Predicting Startup Funding Using AI and CFA

Author: Yan Katcharovski

Published: November 7, 2024

CCSBE 2024 Conference 

Abstract:

Using 600 verbal startup pitches, we explore the application of Artificial Intelligence (AI) and the Critical Factor Assessment (CFA) to predict business angel funding outcomes. Contrary to prior research focusing on features derived from pitch metadata and voice recordings, we employ a Large Language Model (LLM) to synthesize key CFA factors, which assess startups across eight critical dimensions. These synthesized features are then used in machine learning models, such as Naive Bayes and Support Vector Machines, to predict investment decisions. Our findings reveal that specific factor combinations such as Features & Benefits, Readiness, and Financial Expectations consistently result in improved predictive capabilities, achieving accuracy of 79%, F1 score of 0.85 and average precision of 0.84. We demonstrate that models using synthesized CFA factors can effectively screen investment opportunities, offering a scalable and cost-effective method for early-stage funding evaluations.