JPMorgan Chase plans to introduce a new generation of artificial intelligence agents later this year that can operate on their own for hours, marking a notable step in how large companies are using AI, according to a CNBC report.
The bank's chief analytics officer, Derek Waldron, said the technology is moving beyond simple prompts and single tasks. Instead of acting like narrowly focused tools, the next wave of AI agents is designed to carry out multi-step workflows across different software systems with less human oversight.
Waldron said the industry has entered what he described as the era of long-running autonomous agents. In practical terms, that means systems that can stay active for an hour or two while completing a goal rather than needing constant human direction after a few minutes.
The shift matters because it suggests AI is becoming capable of more complex enterprise work. Waldron said improvements in reasoning have made these systems better at handling what he called intellectual coherence, or the ability to remain on task long enough to be useful in business settings.
He compared the emerging systems to team managers rather than individual workers. In that view, an AI agent can break a problem into parts, assign steps to different tools, and keep the work moving for longer periods.
Recent technical advances have also helped. Waldron pointed to AI systems that can write code, navigate web browsers and interact with desktop software, all of which broaden what agents can do inside corporate environments.
JPMorgan does not yet consider the technology ready for full-scale deployment because of security and governance concerns. Still, Waldron said the bank expects to have those capabilities in 2026, suggesting the firm sees the barrier to adoption as temporary rather than fundamental.
The bank is already seeing business gains from AI, especially in private banking. Waldron said AI tools are used overnight to review market activity, client positions and research so bankers can spend more of their time with customers.
According to Waldron, those tools have helped drive a 20% increase in gross sales in private banking. He added that the technology could eventually allow each banker to cover as many as 50% more clients.
That potential is especially significant for JPMorgan, the largest U.S. bank by assets and one of the biggest corporate spenders on technology, with an annual tech budget of nearly $20 billion.
Much of the early conversation around generative AI has focused on reducing expenses and automating routine work. Waldron said the bank now sees a broader opportunity. In his view, enterprises that succeed with AI will not simply eliminate the most jobs. They will use the technology to build a durable competitive advantage and grow revenue.
JPMorgan chief executive Jamie Dimon has said some employees will be displaced by AI, but the bank plans to retrain and redeploy workers affected by the shift.
Waldron also said AI is changing the economics of enterprise software. As JPMorgan evaluates whether to buy technology from vendors or build it internally, he said the bank is increasingly willing to develop more capabilities in-house. That trend, he said, could weaken the advantage of some software suppliers.
The bank's move reflects a wider change in corporate AI strategy. Companies are no longer just asking whether AI can answer questions or draft text. They are now testing whether it can stay coherent long enough to complete meaningful work across entire business processes.