In an effort to address the significant challenge of modernizing outdated legacy code, Morgan Stanley has developed an in-house artificial intelligence solution designed specifically for this purpose. This innovative tool aims to facilitate the conversion of legacy programming languages into contemporary, more efficient coding languages.
Launched in January, the AI system, named DevGen.AI, was introduced as a response to the advancements in OpenAI’s GPT models, according to The Wall Street Journal. DevGen.AI specializes in translating older programming scripts-such as Perl, which was first released in 1987-into plain English. Developers can then interpret this simplified output to rewrite the code in modern languages like Python, streamlining the modernization process.
Transforming Legacy Code: A Strategic Move
Mike Pizzi, Morgan Stanley’s global head of technology and operations, revealed to WSJ that over the first five months since its deployment, DevGen.AI has processed approximately 9 million lines of code. This effort has resulted in saving the firm’s 15,000 software engineers around 280,000 hours of manual labor, significantly accelerating their development cycles.
Why Build In-House? The Limitations of Commercial Solutions
Pizzi explained that Morgan Stanley chose to develop this tool internally because existing commercial AI products did not fully meet their specific needs. Many off-the-shelf solutions lacked the capability to decode proprietary or highly specialized coding languages unique to the financial institution, making an in-house approach more viable.
“Creating our own solution provided us with unique functionalities that are not available in the current market offerings,” Pizzi stated. “This approach allowed us to stay ahead of the curve and innovate early.”
Current Capabilities and Future Prospects
DevGen.AI has been trained on a variety of programming languages, including those tailored specifically for Morgan Stanley’s operations. While the AI can perform full translations from legacy to modern languages, it still lacks the nuanced understanding and creative problem-solving skills of human developers. Consequently, Morgan Stanley continues to rely on skilled engineers to oversee and refine the translation process.
Despite the AI’s capabilities, the firm has not reduced its software engineering team solely because of this technology. Although Morgan Stanley laid off 2,000 employees from its 80,000-strong workforce in March, Pizzi emphasized that the AI tool is viewed as an augmentation rather than a replacement for human expertise.
Broader AI Initiatives at Morgan Stanley
Beyond code modernization, Morgan Stanley has introduced several AI-driven applications to enhance employee productivity. These include tools that summarize lengthy video meetings and systems that swiftly extract relevant information from the firm’s extensive research database.
Leadership’s Vision for AI Integration
Last year, Morgan Stanley’s CEO Ted Pick highlighted the transformative potential of AI tools, suggesting they could free up to 15 hours per employee each week. He described these innovations as potentially “game-changing,” underscoring the firm’s commitment to leveraging AI for operational efficiency, as reported by Reuters.
