In its 2026 company update, JPMorganChase outlined expected technology expenses of approximately US$19.8 billion. This covers the bank’s technology portfolio, not a budget dedicated exclusively to artificial intelligence. That distinction matters before using the figure as a reference for another company.
Scale is not an investment plan
A smaller organization does not need to reproduce the budget or architecture of a global bank. It needs to understand the constraint it is trying to address. Modernization, operational continuity, new products and experimentation compete for resources, but they should not be assessed as though they deliver the same kind of benefit. Copying the figure without identifying that difference does not help with prioritization.
Start with the process and its owner
Choose a workflow with an observable problem and someone accountable for the outcome. Record workload, recurring errors, manual interventions and response times before making a change. This baseline makes it possible to discuss a concrete hypothesis. It also prevents a model demonstration from turning into a deployment commitment before anyone knows who will look after the operation.
Separate investment, activity and outcomes
Buying a platform, integrating an API and releasing a feature are delivery milestones. On their own, they are not evidence of returns. Define in advance which indicators support continuing, adjusting or stopping the work. Include integration, human review and maintenance costs. Compare the same process and state the observation period and its limitations.
Authorize a verifiable increment
The next step might be a small integration, a data correction or a technical assessment. Choose the delivery that clarifies the most important decision, with bounded exposure and a way to reverse it. Another organization’s large budget does not prove that Kubernetes, a proprietary platform or a particular model is the right choice for your context.
Next step
F.A.L A.I Agency proposes starting with a fit assessment of the context and decision. Where there is a fit, the executive diagnosis is a separate paid engagement. The aim is to define the next delivery and the evidence needed to evaluate it.