JPMorgan Chase plans to deploy more powerful AI agents this year

JPMorgan Chase is gearing up to introduce cutting-edge artificial intelligence agents later this year, designed to operate autonomously for prolonged periods, according to exclusive information obtained by CNBC. This step signifies a notable advancement in the corporate realm's integration of AI technology. Derek Waldron, the bank's chief analytics officer, explained in an interview that these AI agents are evolving from simple task performers into sophisticated digital employees capable of managing workflows across various software applications. "We have entered the era of long-running autonomous agents," Waldron stated, highlighting their ability to function for durations of one to two hours, rather than just a few minutes. The emergence of long-running agents over the past year, exemplified by tools like Anthropic's Claude Code and OpenClaw, showcases the potential of this technology. JPMorgan's initiative indicates that it is nearing the resolution of security and governance challenges that have hampered widespread adoption in large enterprises. As the largest bank in the U.S. by assets, with a technology budget nearing $20 billion, JPMorgan is positioned to lead in this space. Waldron emphasized that while discussions around generative AI often focus on the intelligence of models, a critical consideration is how long these systems can operate efficiently before necessitating human oversight. He referred to this idea as "intellectual coherence," which has been enhanced by advancements in AI reasoning capabilities. This allows AI agents to function more as team managers capable of delegating tasks, akin to human supervisors. Recent developments have enabled these agents to undertake more complex responsibilities, such as coding, browsing the web, and interacting directly with desktop applications. Although these long-running agents are not yet ready for corporate deployment due to security issues, Waldron anticipates their introduction by 2026, with the potential for them to maintain coherence for extended periods. The impact of AI on productivity has been particularly evident in software development and back-office operations, but Waldron noted its growing influence on revenue-generating roles as well. For instance, in private banking, AI systems analyze market conditions and client portfolios overnight, allowing bankers to concentrate on client engagement. This has resulted in a reported 20% increase in gross sales, with potential for individual bankers to broaden client outreach by up to 50%. While CEO Jamie Dimon acknowledged that AI may displace some employees, he assured that the company is committed to retraining and redeploying those affected. Waldron added that companies are shifting their perspective on AI from a cost-cutting mechanism to a tool for revenue expansion. "For enterprises to succeed with AI, it’s not just about reducing headcount; it’s about creating a sustainable competitive edge," he stated. Moreover, JPMorgan's strategy regarding software development is evolving. The firm is increasingly evaluating its ability to develop capabilities internally, which may challenge traditional software vendors. Waldron noted, "The barriers protecting certain software companies are certainly lowering compared to previous times."