Deep within the bowels of JPMorgan Chase’s data centers and cloud providers, an artificial intelligence program crucial to the bank’s aspirations grows more powerful by the week.

The program, called LLM Suite, is a portal created by the bank to harness large language models from the world’s leading AI startups. It currently uses models from OpenAI and Anthropic.

Every eight weeks, LLM Suite is updated as the bank feeds it more from the vast databases and software applications of its major businesses, giving the platform more abilities, Derek Waldron, JPMorgan chief data analytics officer, told CNBC in an exclusive interview.

“The broad vision that we’re working towards is one where the JPMorgan Chase of the future is going to be a fully AI-connected enterprise,” Waldron said.

JPMorgan, the world’s largest bank by market capitalization, is being “fundamentally rewired” for the coming AI era, according to Waldron. The bank, a heavyweight across Main Street and Wall Street finance, wants to provide every employee with AI agents, automate every behind-the-scenes process and have every client experience curated with AI concierges.

If the effort succeeds, the project could have profound implications for the bank’s employees, customers and shareholders — even the nature of corporate labor itself.

Waldron, who gave CNBC the first demonstration of its AI platform seen by any outsider, showed the program creating an investment banking deck in about 30 seconds, work that would’ve previously taken a team of junior bankers hours to complete.

Since the arrival of OpenAI’s ChatGPT in late 2022, optimism over generative AI has driven markets higher on gains from the tech giants and chip makers closest to the trade. Underpinning their growth is the expectation that corporate clients deploying AI will either boost worker productivity or lower expenses through layoffs — or both.

But similar to how the internet story played out in the 1990s, near-term expectations for AI may have outstripped reality. Most corporations had no tangible returns yet on their AI projects despite more than $30 billion in collective investments, according to an MIT report from July.

In the case of JPMorgan, even with it $18 billion annual tech budget, it will take years for the company to realize AI’s potential by stitching the cognitive power of AI models together with the bank’s proprietary data and software programs, said Waldron.

“There is a value gap between what the technology is capable of and the ability to fully capture that within an enterprise,” Waldron said.

Companies “do work in thousands of different applications, there’s a lot of work to connect those applications into an AI ecosystem and make them consumable,” he said.

If JPMorgan can beat other banks to the punch on incorporating AI, it will enjoy a period of higher margins before the rest of the industry catches up. That first-mover advantage will allow it to grow revenues faster by going after a larger slice of the addressable market in global finance — enabling the bank to pitch more middle-market companies in investment banking, for instance.

AI was a major topic at a four-day executive retreat held in July by JPMorgan CEO Jamie Dimon, according to a person who attended but declined to be identified speaking about the private event.

Among concerns discussed at the off-site meeting, held at a resort outside Nashville, was how AI-driven changes will be adopted by the bank’s 317,000-person workforce and its possible impacts to the apprenticeship model on areas including investment banking.

If JPMorgan succeeds with its AI goals, it will mean that a bank that is already the largest and most profitable in American history is set for new heights. Dimon has led the bank since 2005, guiding it through periods of upheaval to notch record profits in 7 of the last 10 years.

The end state for JPMorgan, as envisioned by Waldron, is a future in which AI is woven into the fabric of the company:

“Every employee will have their own personalized AI assistant; every process is powered by AI agents, and every client experience has an AI concierge,” he said.

JPMorgan laid the groundwork for this starting in 2023, when it gave employees access to OpenAI’s models through LLM Suite; it was essentially a corporate ChatGPT tool used to draft emails and summarize documents.

About 250,000 JPMorgan employees have access to the platform today, which is the entire workforce except for branch and call center staff, said Waldron. Half of them use it roughly every day, he said.

JPMorgan is now early in the next phase of its AI blueprint: It has begun deploying agentic AI to handle complex multistep tasks for employees, according to an internal roadmap provided by the bank.

“As those agents become increasingly powerful in terms of their AI capabilities and increasingly connected into JPMorgan,” Waldron said, “they can take on more and more responsibilities.”

He gave the program a prompt: “You are a technology banker at JPMorgan Chase preparing for a meeting with the CEO and CFO of Nvidia. Prepare a five-page presentation that includes the latest news, earnings and a peer comparison.”

LLM Suite created a credible-looking PowerPoint deck in about 30 seconds.

“You can imagine in the past how that would have been done; we would’ve had teams of investment banking analysts working long hours at night to do this,” said Waldron.

The bank is also training AI to draft other key investment banking documents including the “inch thick” confidential memos that JPMorgan produces for prospective M&A clients, said the person who attended the July executive meeting.

The prospect of collapsing work loads means that fewer junior bankers may be needed even while AI-enabled teams handle more work and pitch more companies, according to senior Wall Street executives at several firms who spoke on the condition of anonymity to provide their candid thoughts.

But to extract the full value from this new, almost magical technology, it’s not just about the tools: Changes to how employees and departments are organized may be needed.

One proposal being discussed at a major investment bank is reducing the ratio of junior bankers to senior managers from the current 6-1 to 4-1. In the new regime, half of those junior bankers would be working from cities with cheaper labor, say Bengaluru, India, and Buenos Aires, Argentina, instead of being clustered in expensive New York.

The AI-powered junior bankers could then work on deals in shifts around-the-clock, passing the baton from one time zone to the next.

With fewer bankers on the payroll, the cost structure of investment banking would fall, boosting the bottom line, said the executives.

In May, JPMorgan’s consumer banking chief told investors that operations staff would fall by at least 10% in the next five years thanks to AI deployment.

“In an AI world, you’ll still have people at the top who are managing and have relationships with clients, but many, many of the processes underneath are now being done by AI systems,” Waldron said.

Despite market concerns that the AI trade is a brewing bubble, corporate clients are actually more worried now that if they don’t start adopting it soon, they’ll fall behind and lose share, said Avi Gesser, a Debevoise & Plimpton partner who advises corporations on issues around AI.

“People are starting to see what these tools can do,” Gesser said. “They’re sort of like, ‘Wow, if you get the workflow right, implement it properly and have the right guardrails, I could see how that would save you a lot of time and a lot of money and deliver a better product.”