The Impact of Artificial Intelligence on Investment Banking

Authors:

Abstract
We investigate how artificial intelligence (AI), including generative AI (GenAI) and emerging agentic workflows, may reshape investment banking from 2026 to 2036. The central problem is that public claims about AI in banking tend to outpace the actual evidence of adoption in advisory, valuation, diligence, and client execution work. We synthesise research from four related literatures: GenAI productivity in information-intensive work, task-based labour economics, finance scholarship on data technologies and market structure, and governance research on automation bias, hallucination, and accountability. We use a claim-evidence framework that treats prior AI-impact claims as hypotheses rather than proof. Each claim is assessed using evidence strength, domain fit, and plausible adoption constraints. The report finds that the strongest evidence supports substantial augmentation of documentheavy and synthesis-heavy junior work, rather than the full replacement of investment banking judgment. Competitive advantage is likely to shift toward proprietary data, client access, governance capacity, and decision-making under uncertainty. The main risk is not only technical failure; it is overreliance on fluent but unreliable outputs in high-stakes deal contexts.