American Express cuts 5% of workforce in AI-driven management restructuring
American Express is eliminating approximately 5% of its global workforce — around 3,500 roles — as part of a restructuring that CFO Christophe Le Caillec framed as preparation for an AI-enabled operating model. The company booked a $336 million restructuring charge alongside a third-quarter profit that beat analyst expectations.

The restructuring's explicit framing around AI readiness and management-layer flattening signals a structural shift in how a major card issuer is repositioning its operating model, with implications for the wider financial-services industry.
American Express to Cut Approximately 5% of Global Workforce in AI-Driven Restructuring
American Express has announced a reduction of approximately 5% of its global workforce, affecting roughly 3,500 employees, as the company restructures its management layers and repositions for an AI-driven operating model. CEO Stephen Squeri disclosed the cuts on the company's third-quarter earnings call, alongside a $336 million restructuring charge tied to the programme.
The announcement came as Amex reported third-quarter profit that beat analyst expectations, lending the restructuring a tone of strategic offence rather than financial distress.
What the Company Said
CFO Christophe Le Caillec provided the clearest statement of intent, characterising the restructuring as an effort to make American Express "leaner, more agile, and also to prepare for AI," according to reporting by HR Reporter and CNBC independently. The programme involves flattening management layers and increasing spans of control — structural changes consistent with organisations seeking to reduce overhead ahead of broader automation deployment.
Amex's CFO framed the restructuring not as a cost-cutting reaction but as deliberate preparation for an AI-enabled operating model, according to HR Reporter and CNBC.
The $336 million restructuring charge is a verified material cost tied directly to the workforce reduction. The company has disclosed this figure in connection with the cuts; no independent verification from SEC filings was possible for this report, as those documents were not accessible at time of publication.
Scale and Scope
The approximately 5% reduction translates to roughly 3,500 roles across the global business, corroborated by multiple outlets including LiveMint, Payments Dive and Banking Dive. The precise total employee baseline carries some uncertainty: LiveMint cited a figure of approximately 78,000 employees, while HR Reporter referenced almost 77,000 at the end of 2025. The Fin Desk is not treating either figure as definitive; what is confirmed is the approximately 5% proportional cut and the approximately 3,500 absolute headcount figure.
The restructuring's focus on flattening management and widening spans of control suggests the affected roles are concentrated in middle-management and supervisory layers, though the company has not publicly specified which functions or geographies bear the heaviest impact.
Payments Industry Context
The following represents editorial analysis and should be read as such.
Analysts covering the payments sector have noted that card networks and issuers with mature technology estates have been under pressure to demonstrate that prior investment in AI and machine-learning capabilities translates into measurable efficiency gains. Restructurings framed explicitly around AI readiness — as Amex's CFO has done here — are increasingly common across financial services, though the scale and the accompanying management-layer flattening give this programme a more structural character than a typical headcount reduction.
American Banker's headline coverage noted Amex is simultaneously "touting AI" and planning the downsizing, a pairing that reflects a wider industry tension: institutions must invest in new capabilities while demonstrating to investors that those investments ultimately reduce unit costs. The $336 million charge is a near-term earnings headwind, but the Q3 profit beat suggests investors are not interpreting the restructuring as a sign of underlying weakness.
Why It Matters
For the payments and fintech industry, the Amex announcement carries several implications worth watching:
- AI as workforce rationale: This is among the more explicit deployments of AI preparedness as a stated driver of large-scale restructuring by a major payments network, as reported by HR Reporter and CNBC.
- Management architecture shift: The emphasis on span-of-control expansion and layer reduction suggests a structural, not merely tactical, reorganisation.
- Earnings context: The Q3 profit beat, confirmed by HR Reporter and Latestly, reduces the risk that markets interpret the restructuring charge as distress signalling.
- Restructuring cost: The $336 million charge is material and verified; its full impact on full-year earnings will depend on timing of recognition.
Editors' note: Source articles in the research package carry a publication date of October 23, 2026. The Fin Desk is seeking confirmation of the underlying earnings call date before finalising datelines. This article will be updated when that verification is complete.
The Fin Desk Newsroom publishes verified reporting on the developments shaping fintech, payments and modern financial infrastructure.
Related Stories

Visa Reports FY2026 Results After Month of Moves Spanning Stablecoins, AI and China
Visa closed its fiscal fourth quarter on 27 October 2026 with full-year results, capping a month that included the production launch of stablecoin-powered money movement on Visa Direct, a live AI agent payments pilot with BILL and Billtrust, and regulatory approval to begin preparations for a domestic business in China.

Mastercard, Corpay, Ramp and rivals launch purpose-built cards for AI agents
A cluster of product launches from Mastercard, Corpay, Sunrate, AppZen and Ramp between 2025 and 2026 is consolidating around payment cards designed to operate under AI agent control rather than human cardholders. The trend spans card network, issuer and software layers, signalling that autonomous software spend is becoming a payments infrastructure question.

Visa Benchmark Finds AI Agents Fail 42.8% of Shopping Scenarios as Agentic Commerce Push Accelerates
Visa has published benchmark data showing AI models failing in more than four in ten simulated buying scenarios, even as the company moves agentic B2B payments into live pilots with BILL and Billtrust and shifts stablecoin-powered money movement into production.

Visa launches Commercial Integrated Partners programme to embed payment APIs with fintechs
Visa launched a structured partner programme on 21 May 2025 that gives fintechs advanced APIs to embed Visa commercial products directly into their applications. Completed integrations are then made available to Visa's financial-institution clients, with Car IQ named as one of the first participants.
The essential developments in modern finance
The essential developments across fintech, payments and modern finance — delivered to your inbox.
Free. No spam. Unsubscribe anytime.