AmEx Cuts 5% of Staff Weeks After OCC Flags AML Gaps
American Express has announced it will lay off roughly 5% of its global workforce as part of a $336 million restructuring, and that headline alone would be unremarkable corporate news. What makes it significant for compliance professionals and digital asset accounting teams is the timing: the cuts come just 15 days after the US Office of the Comptroller of the Currency publicly cited AmEx for maintaining "inadequate resources, including staff without sufficient expertise" inside its anti-money laundering program. Add a separate $350 million penalty for missing $13 billion in suspected money laundering activity, and the picture that emerges is one of a major financial institution shrinking its headcount at precisely the moment its regulator has told it to invest more, not less, in AML infrastructure.
What AmEx Actually Announced
AmEx's Q3 2026 earnings report contained two distinct stories running in parallel. On the financial side, the company raised the floor of its full-year adjusted earnings-per-share guidance from $17.30 to $17.60, leaving the $17.90 ceiling unchanged. That shifted the midpoint of the range from $17.60 to $17.75, a gain of roughly 0.85%. By any measure, a forecast upgrade of less than 1% is modest, and investors responded by selling the stock down 3.7% by midday in New York. Year-to-date, the shares have lost 15% of their value.
The restructuring math
The $336 million restructuring charge is more than offset by a one-time pretax gain of nearly $1 billion from the divestiture of AmEx's stake in Global Business Travel. CEO Stephen Squeri noted on the earnings call that the company is "dropping two thirds of the gain to the bottom line." CFO Christophe Le Caillec told CNBC the workforce reduction is designed to make AmEx "leaner, more agile" and positioned to "automate and take advantage of AI." The explicit operational logic is to reduce management layers so that fewer managers oversee more staff, freeing budget for AI tooling.
Scale of the cuts
At approximately 5% of total headcount, the reduction represents a material organisational change. AmEx has not disclosed which departments will bear the cuts, and nothing in its public earnings materials clarifies whether compliance, financial crime, or AML functions are in scope. That silence is the crux of the concern.
The OCC Action: What the Regulator Actually Said
On 8 October 2026, the OCC issued a finding against AmEx specifically targeting the resourcing of its AML program. Comptroller Jonathan Gould stated that banks must "devote sufficient resources to ensure compliance with laws and regulations designed to detect and prevent money laundering." The OCC's language centred on two deficiencies: an overall resource shortfall and a skills gap, meaning the staff AmEx did have in compliance roles lacked sufficient AML expertise.
The $350 million penalty context
Separately from the October OCC action, AmEx is contending with a $350 million penalty tied to failures to identify approximately $13 billion in suspected money laundering activity. That figure represents a systemic gap in transaction monitoring and suspicious activity reporting, not an isolated incident. When a regulator cites inadequate staffing and expertise in the same period that a nine-figure penalty is being processed, the two data points reinforce each other. The OCC is, in effect, describing the organisational conditions that allowed a $13 billion monitoring gap to persist.
Why the Timing Creates a Compliance Paradox
This is the tension that accounting firms, CFOs, and financial crime officers need to sit with. A regulator has told a large US bank that it has too few qualified people doing AML work. Fifteen days later, that bank announces it is reducing its overall workforce by 5%. The company has not confirmed that compliance will be protected, and the restructuring rationale, centred on AI automation and flatter management structures, does not inherently exclude financial crime functions.
AI as a compliance argument
AmEx's management is framing AI as the reason headcount reductions are operationally safe. That argument has some validity in narrow contexts: machine learning models can flag transaction anomalies faster than manual review, and natural language processing can assist with SAR narrative drafting. But the OCC's criticism was not primarily about throughput. It was about expertise. Automated systems require skilled human oversight to tune detection thresholds, investigate alerts, make materiality judgements, and file accurate Suspicious Activity Reports. Reducing the number of experienced AML professionals while simultaneously facing a regulator who says you already have too few is a difficult position to defend in a future examination.
Span-of-control risk in financial crime
The stated plan to "reduce layers and increase spans of controls" is a standard efficiency model in sales, technology, and operations. Applied to AML, it carries specific risk. A compliance officer managing a larger caseload of alerts and a broader team is more likely to miss nuanced patterns, approve borderline cases without adequate review, and fail to document decision rationale at the level regulators expect. Span-of-control expansion in financial crime functions is not the same productivity trade-off it is in, say, a customer service organisation.
Accounting and Reporting Implications for B2B Readers
For accounting firms advising financial services clients, and for CFOs at institutions with AML obligations, the AmEx situation surfaces several practical considerations worth tracking in your own governance frameworks.
Restructuring charges and one-time gains
AmEx's decision to fund its $336 million restructuring charge with a nearly $1 billion asset-disposal gain is straightforward under US GAAP: the divestiture gain is recognised in the period the sale closes, and the restructuring charge is accrued when the plan is communicated and the criteria in ASC 420 are met. Both flow through the income statement. The net effect is a substantial one-time earnings boost that inflates the period's bottom line without reflecting ongoing operational performance. Analysts and audit committees should strip both items when assessing normalised earnings quality, and any forward guidance based on adjusted EPS already excludes them.
Regulatory penalty accruals
The $350 million AML penalty requires careful treatment. Under ASC 450 (Contingencies), a loss contingency must be accrued when it is probable and the amount can be reasonably estimated. If AmEx has reached or is near settlement, the full amount should appear as a liability. Audit teams reviewing financial services clients with outstanding regulatory matters should confirm that AML-related penalties are neither under-accrued nor disclosed only in footnotes when the settlement timeline is concrete.
What this signals for crypto-adjacent financial institutions
The AmEx case is a traditional banking story, but its lessons transfer directly to digital asset businesses. The OCC's language about "staff without sufficient expertise" maps precisely onto the most common AML deficiency finding at crypto exchanges and custodians: transaction monitoring systems that generate alerts that nobody with appropriate training is reviewing. Any institution, whether a bank or a crypto-asset service provider, that is cutting compliance headcount to fund an AI transition should document in writing how it will maintain examination-ready AML coverage during the transition period. For firms that rely on crypto accounting software to track on-chain flows, that software is an input to AML, not a substitute for the human judgement that regulators assess.
Internal audit functions should also consider requesting a formal compliance resourcing attestation from management: a statement that planned headcount reductions have been reviewed against AML staffing obligations and that no regulated function will fall below the minimum coverage needed for the next supervisory examination cycle. If management cannot provide that attestation, that is itself a finding.
What Compliance Officers and Firms Should Do Now
The AmEx situation is a useful prompt for any firm currently planning a workforce reduction or AI-driven efficiency programme. A few concrete steps are worth taking immediately.
Ring-fence your AML headcount in any restructuring plan
Before finalising redundancy lists, compliance leadership should produce a headcount-to-obligation map: for each AML control (transaction monitoring, SAR filing, CDD refresh, correspondent banking review, etc.), document the minimum FTE coverage needed to meet regulatory expectations. That document becomes the evidence base if an examiner later questions whether the restructuring impaired the program.
Review AI deployment against regulatory expectations
If your institution is deploying AI tools to assist AML functions, ensure the model governance documentation covers how humans remain accountable for alert adjudication. The FinCEN guidance on innovation and the OCC's model risk management framework (OCC Bulletin 2011-12, updated via subsequent interagency guidance) both require that automated systems be validated and that human review be documented. Relying on AI without auditable human decision points will not satisfy an examiner, regardless of the technology's accuracy rate.
Disclose restructuring impacts on compliance in board reporting
CFOs and chief compliance officers should ensure that board and audit committee materials explicitly address whether any planned restructuring affects regulated compliance functions. That disclosure protects the board, satisfies governance best practice, and creates a contemporaneous record that leadership considered the regulatory risk before proceeding.
For firms using digital asset accounting software or crypto bookkeeping software to manage on-chain transaction records, ensure those systems are integrated into your broader AML workflow documentation. An isolated tool that produces clean records but is not connected to your SAR process or KYC refresh cycle will not satisfy an examiner who is looking at program-level adequacy, not just data quality. See also our earlier analysis of FinCEN's rising fraud warnings at crypto kiosks, which covers how transaction monitoring gaps at crypto points of service attract exactly the kind of OCC-style scrutiny AmEx is now facing. For a broader view of how enforcement actions intersect with compliance resourcing, the AMLA AML/CFT guidance for CASPs under MiCAR provides a useful parallel from the European regulatory tradition.
Frequently Asked Questions
What did the OCC specifically find wrong with AmEx's AML program?
The OCC stated on 8 October 2026 that AmEx had "inadequate resources, including staff without sufficient expertise" in its anti-money laundering program. Comptroller Jonathan Gould emphasised that banks must devote sufficient resources to comply with laws designed to detect and prevent money laundering. The OCC did not specify which business lines or geographies were most affected.
Is the $350 million penalty connected to the OCC's October finding?
The source material treats these as separate matters. The $350 million penalty relates to a failure to identify approximately $13 billion in suspected money laundering activity. The October OCC action focuses on resource adequacy. Both concern AML program quality, and both are active at the same time, but they appear to stem from distinct regulatory processes.
How should a CFO account for a $336 million restructuring charge funded by an asset-disposal gain?
Under US GAAP, the two items are recorded separately. The disposal gain is recognised when the transaction closes; the restructuring charge is accrued under ASC 420 when the plan meets the required criteria. Both flow through the income statement in the relevant period. Adjusted EPS figures typically exclude one-time items of both types, so management's guidance figures will already strip them out. Audit committees should request a bridge between GAAP and adjusted earnings to ensure the stripping is applied consistently.
Can AI tools replace human AML analysts in a restructuring?
Not in a way that satisfies current US regulatory expectations. The OCC's model risk management framework requires human accountability for automated decisions, and FinCEN's rules on suspicious activity reporting require that a qualified individual review and sign off on SARs. AI can assist with alert triage and pattern detection, but it does not remove the legal obligation to employ people with sufficient AML expertise to make and document materiality judgements.
What should accounting firms advising financial services clients take from this case?
Three things: first, check that any restructuring your client is planning includes a documented compliance resourcing review before headcount decisions are finalised. Second, confirm that any pending AML-related regulatory penalties are correctly accrued under ASC 450 and not merely disclosed in footnotes. Third, if AI tooling is being deployed in the compliance function, verify that model governance documentation and human review processes are in place and auditable.
Source: Protos