AI & Agentic AI in SME Credit

AI-native underwriting has rapidly moved from pilot to production, delivering measurable gains in cost-to-serve and expanding access to credit for previously underserved SMEs. However, as autonomy increases, accountability has struggled to keep pace. With the emergence of agentic AI, the ability to fully trace and explain individual credit decisions is becoming more complex. At the same time, models trained on historical data risk reinforcing structural biases—systematically favoring urban, formally registered, and male-led enterprises and do not sufficiently take into account forward looking variables.

This session will explore where AI-native underwriting is operating at scale, what the evidence shows on cost efficiency and portfolio performance, and how governance frameworks can evolve to ensure that greater automation is matched by robust standards on fairness, transparency, and explainability.