JPMorgan Chase, the largest U.S. bank, is moving this year to introduce AI agents that can perform tasks autonomously for far longer than before. Beyond automating simple work, the concept of “digital employees” that handle multi-step tasks on their own and move across different software is expected to be applied in finance in earnest.

CNBC reported on June 9 that JPMorgan is pushing to deploy AI agents capable of processing work autonomously for extended periods, beyond the level of performing a single task.

Derek Waldron (데릭 월드런), JPMorgan’s chief analytics officer, said in a recent interview that AI agent technology is entering the stage of “long-duration agents.” He said that if existing AI agents carried out tasks for 2 to 3 minutes based on human instructions, they are now advancing toward a level that can pursue goals for 1 to 2 hours.

This shows that the focus of the generative AI race is shifting from simple model performance to how long it can reliably handle real work.

JPMorgan said long-duration autonomous AI is not yet ready for immediate deployment in the field, but that technology usable in corporate environments is quickly approaching. Waldron forecast that “security and governance issues still exist, but in 2026 we will have this kind of AI agent.”

JPMorgan’s move is seen as an example showing that preparations are accelerating to address security and control issues that have hindered big companies’ adoption of AI. JPMorgan, which invests $20 billion a year in technology, is focusing on developing AI not as a simple work-assistance tool but as digital manpower that directly manages workflow.

Waldron stressed that the core of AI competitiveness lies less in the model’s intelligence itself than in how reliably it can perform work without human intervention. He called this “intellectual consistency” and said the latest AI agents are beginning to play a role closer to a team manager than an individual practitioner. He said AI is evolving to break down tasks, distribute work to the right tools and manage the overall process.

Technical progress is also happening quickly. Recent AI agents can write code and also operate web browsers and control desktop applications, gaining the ability to perform more complex tasks. As a result, productivity gains are expanding beyond development departments or back offices into areas that generate revenue.

A representative case is JPMorgan’s personal wealth management division. There, an AI system automatically analyses overnight market moves, changes in client positions and research materials, and provides the results to employees.

This has allowed employees to focus on client consultations rather than collecting materials, and Waldron said revenue in the division rose about 20 percent through such AI use. He also forecast the number of clients each employee can handle could increase by up to 50 percent.

Workforce structure changes stemming from the spread of AI are also drawing attention. Jamie Dimon (제이미 다이먼), JPMorgan’s CEO, has already mentioned that some jobs could be replaced by AI. The company is known to be preparing retraining and job reassignment programmes for affected employees.

Waldron also stressed that companies have moved past the stage of viewing AI simply as a cost-cutting tool. “The core of the AI race is not eliminating as many jobs as possible but securing a sustainable competitive advantage,” he said, adding that companies recently are showing greater interest in revenue expansion effects than cost cutting.

AI advances are also changing software procurement strategies at large financial firms. JPMorgan said it is first reviewing whether it can use AI to develop functions in-house rather than buying external software. Waldron said this change could lower entry barriers for some traditional software companies compared with the past.

The industry views JPMorgan’s plan to introduce AI agents as a case that could show the next stage of AI use in finance. Assessments say autonomous AI that performs real work, beyond simple assistance tools, has begun to change both the financial industry’s work structure and its technology investment strategy.