In the dynamic and data-driven landscape of financial markets, this paper introduces MarketSenseAI, a novel AI-driven framework leveraging the advanced reasoning capabilities of GPT-4 for scalable stock selection. MarketSenseAI incorporates Chain of Thought and In-Context Learning methodologies to analyze a wide array of data sources, including market price dynamics, financial news, company fundamentals, and macroeconomic reports emulating the decision making process of prominent financial investment teams. The development, implementation, and empirical validation of MarketSenseAI are detailed, with a focus on its ability to provide actionable investment signals (buy, hold, sell) backed by cogent explanations. A notable aspect of this study is the use of GPT-4 not only as a predictive tool but also as an evaluator, revealing the significant impact of the AI-generated explanations on the reliability and acceptance of the suggested investment signals. In an extensive empirical evaluation with S&P 100 stocks, MarketSenseAI outperformed the benchmark index by 13%, achieving returns up to 40%, while maintaining a risk profile comparable to the market. These results demonstrate the efficacy of Large Language Models in complex financial decision-making and mark a significant advancement in the integration of AI into financial analysis and investment strategies. This research contributes to the financial AI field, presenting an innovative approach and underscoring the transformative potential of AI in revolutionizing traditional financial analysis investment methodologies.