JPMorgan Chase AI Strategy: How the Top-Ranked AI Bank Builds and Measures AI

JPMorgan Chase ranks first in the Evident AI Index for banks for the fifth year in a row. It gives most of its workforce access to its in-house AI platform, and its leaders say openly that AI has already cut jobs in some areas. Measurement has not kept pace: in October 2026, the bank said AI benefits are not yet visible in its quarterly results.

The ambition is an “AI-connected enterprise.” Chief Analytics Officer Derek Waldron described the goal this way in a McKinsey interview in October 2025.

In practice, this means central AI connected to team knowledge, firm-wide data and the applications people already use.

The focus is credit, fraud and personalization. The COO letter in the bank’s 2025 Annual Report says AI is delivering measurable value in these three areas.

The portfolio is large but concentrated. On the second-quarter 2026 earnings call, CEO Jamie Dimon put the number of AI use cases at almost 1,000. He said about 50 of them really matter, in areas such as risk, fraud, marketing and hedging.

JPMorgan Chase has not said publicly what it will not prioritize.

The platform is built in-house. LLM Suite is the bank’s own generative AI platform. According to Waldron, nearly a quarter of a million people have access. Use was never mandated: access was phased in as people asked for it.

AI spending sits inside a much larger technology budget. The 2025 Annual Report puts the 2026 technology budget at about $19.8 billion.

The AI share is harder to pin down. Dimon told Bloomberg TV in October 2025 that the bank spends about $2 billion a year on AI. In October 2026, CFO Jeremy Barnum put annual AI expense closer to $700 million, and noted that the spending is spread across the firm.

The bank is open about job losses. The annual report says job reductions will likely be the outcome in certain roles, but that “the goal is not fewer headcount.”

In July 2026, Dimon said the bank had reduced jobs by 30% or 40% in discrete areas. He said most affected people were offered jobs elsewhere, and that the bank is preparing to retrain staff.

Training grows with the user. Waldron describes AI training that starts simple and grows in sophistication, from what models can and cannot do to advanced prompting. An “AI Made Easy” program and shared prompt libraries support it.

Ownership is moving. Chief Data and Analytics Officer Teresa Heitsenrether, who led the AI program from a seat on the operating committee, retires at the end of 2026. Chief Technology Officer Scot Baldry takes over her role and the AI portfolio.

Baldry reports to CIO Lori Beer and does not join the operating committee, according to a July 2026 internal memo reported by Bloomberg and Quartz.

Little about governance is public. The annual report refers to connecting models to well-governed data. Published responsible AI principles and a named AI governance body were not found in the sources reviewed.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: None declared, because use of LLM Suite is opt-in.
• Highest level evidenced: Level 3 (Integrating) is claimed only in aggregate, as measurable value in credit, fraud and personalization. No named workflow with a published measure was found.
• Portfolio design: Training is tiered by sophistication. No role-specific AI targets have been disclosed.

The headline value claim has not been updated. In October 2025, Dimon said AI savings roughly matched the bank’s annual AI spend, and called it the tip of the iceberg. This figure is company-claimed and has not been independently verified.

A year later, the bank stepped back from it. On the third-quarter 2026 earnings call, Barnum said the earlier estimate was accurate as far as it went, but that he is moving away from quantifying AI benefits that way.

He also agreed that the benefits are not yet observable in quarterly earnings.

Adoption is the clearest number. Waldron said in October 2025 that a little under half of employees use generative AI tools every day.

No roadmap horizons have been published. The sources reviewed contain no dated 90-day, one-year or multi-year AI milestones.

Change log

• October 2025: Dimon says AI savings match about $2 billion of annual AI spend.
• February 2026: Guy Halamish becomes COO of the commercial and investment bank, with a mandate to speed up AI rollout.
• April 2026: The annual report sets a 2026 technology budget of about $19.8 billion.
• July 2026: Dimon discloses job cuts of 30% to 40% in discrete areas, and AI leadership moves under the CTO.
• October 2026: Evident ranks the bank first for a fifth year. The CFO restates AI spend at about $700 million and says benefits are not yet visible in quarterly results.

• Disclosure: 4 of 4. JPMorgan Chase publicly addresses all four questions of the AI strategy framework. Its answer on measurement is the weakest.
• Scale is proven. Nearly a quarter of a million people can use LLM Suite, and close to half of employees use generative AI daily.
• The workforce effect is real and disclosed. The bank names job cuts of 30% to 40% in some areas, while saying its goal is growth rather than fewer people. Firm-wide headcount rose just 0.4% in 2025.
• The say-do gap is in measurement. The 2025 value claim has not been updated, and the CFO says benefits are not yet visible in results.
• Watch the leadership change. At the end of 2026, AI ownership moves from an operating committee seat into the technology organization.

What the numbers show: Headcount growth has almost stopped: up 0.4% in 2025 while revenue rose about 3%. Dimon has said AI cut jobs by 30% to 40% in some areas, but the firm-wide number has not fallen.

Revenue, headcount and revenue per employee, FY2022 to FY2025

• FY2022: $128.7bn revenue, 293,723 employees, $438k per employee
• FY2023: $158.1bn revenue, 309,926 employees, $510k per employee
• FY2024: $177.6bn revenue, 317,233 employees, $560k per employee
• FY2025: $182.4bn revenue, 318,512 employees, $573k per employee

Change, FY2022 to FY2025: revenue up 42%, headcount up 8%, revenue per employee up 31%.

Notes: Revenue is total net revenue, as reported. Headcount is at year-end. FY2023 onward includes First Republic, acquired in May 2023.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: JPMorgan Chase 10-K filings.

What is JPMorgan Chase’s AI strategy?

JPMorgan Chase aims to become an AI-connected enterprise, where central AI is linked to the bank’s data, knowledge and applications. It concentrates on about 50 high-value use cases in areas such as credit, fraud, risk and personalization, delivered through its in-house platform, LLM Suite.

‍

How many JPMorgan Chase employees use AI?

In October 2025, the bank said nearly a quarter of a million people had access to LLM Suite, and a little under half of employees used generative AI tools every day. Use is voluntary.

‍

Is JPMorgan Chase cutting jobs because of AI?

In some areas, yes. In July 2026, CEO Jamie Dimon said the bank had reduced jobs by 30% or 40% in discrete areas, with most affected people offered jobs elsewhere. The bank says its goal is growth rather than fewer people.

‍

How much does JPMorgan Chase spend on AI?

Its 2026 technology budget is about $19.8 billion. The AI share was given as about $2 billion a year in October 2025 and about $700 million a year in October 2026, and the bank says AI spending is hard to isolate.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings calls and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.