Brex has reportedly begun routing its internal artificial intelligence governance through network-level monitoring, according to CryptoBriefing. The company is said to be using an open-source proxy built for large language models to observe AI traffic across its systems.
A proxy in this context sits between employees or applications and the AI models they call. It can log requests, flag unusual patterns, and apply rules before data reaches an external model. Using an open-source tool means Brex can inspect and modify the underlying code rather than depend on a closed vendor system.
This marks a shift from governance built mainly around written policy and user permissions. Many companies still manage AI use through guidelines telling staff which tools they may use and how. Network-level monitoring instead captures what is actually happening on the wire, regardless of whether policy is followed.
Brex operates in corporate finance, handling expense management, corporate cards, and banking services for businesses. Companies in this space handle sensitive financial data daily. Any AI tool that touches that data raises questions about where information goes and who can see it.
Regulators and enterprise security teams have grown more focused on how firms deploy generative AI internally. Financial companies face particular scrutiny because customer and transaction data can be exposed if an AI tool sends information to an external server. Network traffic monitoring offers a technical checkpoint that does not depend entirely on employees following rules correctly.
The use of an open-source proxy also fits a broader pattern in enterprise software. Companies increasingly prefer tools they can audit directly, rather than trusting a vendor's internal controls without visibility. This is especially relevant for AI, where model behavior can be difficult to predict from the outside.
CryptoBriefing's report does not detail which specific proxy tool Brex is using, nor does it specify the scale of the rollout across the company's infrastructure. It is also unclear whether this change applies to customer-facing products, internal tooling, or both.
The report frames this as part of a wider move within fintech toward treating AI oversight as an infrastructure problem, not just a policy one. Governance built into network monitoring can, in principle, apply uniformly across teams, without requiring every employee to remember a rule.
Market Impact
For now, the reported change appears limited to internal operations at Brex rather than a public product feature. It does not directly affect trading markets or crypto asset prices. Its relevance lies in signaling how fintech firms handling sensitive financial data are adapting internal controls as AI tools become more common in daily workflows.
If other fintech and crypto-adjacent companies adopt similar network-level AI monitoring, it could influence vendor selection and technical architecture across the sector. Firms managing custody, payments, or trading infrastructure may face similar pressure to demonstrate technical, auditable oversight of AI systems, rather than policy alone.
The reported shift at Brex highlights a broader question facing finance-adjacent companies: how to oversee AI tools with the same rigor applied to other sensitive infrastructure. Further detail on the scope and tools involved has not yet been made public.
Frequently Asked Questions
What is an LLM proxy?
An LLM proxy is software that sits between users or applications and a large language model. It can log, filter, or control requests before they reach the AI model.
Why would a fintech company monitor AI traffic at the network level?
Network-level monitoring captures actual AI usage rather than relying on employees to follow written policy. This can help detect when sensitive financial data might be sent to an external AI service.
Does this change affect Brex's customers directly?
The report from CryptoBriefing does not specify whether the monitoring applies to customer-facing products or only internal systems, so the direct impact on customers is unclear.
Is this related to cryptocurrency markets?
The reported change concerns internal AI governance at a fintech company and is not tied to crypto asset prices or trading activity directly.