Fintech Brex is betting that the way forward for enterprise AI isn’t higher orchestration — it’s much less of it.
As generative AI brokers transfer from copilots to autonomous techniques, Brex CTO James Reggio says conventional agent orchestration frameworks have gotten a constraint moderately than an enabler. As a substitute of counting on a central coordinator or inflexible workflows, Brex has constructed what it calls an “Agent Mesh”: a community of slim, role-specific brokers that talk in plain language and function independently — however with full visibility.
“Our purpose is to make use of AI to make Brex successfully disappear,” Reggio instructed VentureBeat. “We’re aiming for whole automation.”
Brex discovered that for its functions, brokers have to work in slim, particular roles to be extra modular, versatile, and auditable.
Reggio mentioned the architectural purpose is to allow each supervisor in an enterprise “to have a single level of contact inside Brex that’s dealing with the totality of their duties, be it spend administration, requesting journey, or approving spend restrict requests.”
The journey from Brex Assistant
The monetary companies trade has lengthy embraced AI and machine studying to deal with the large quantities of knowledge it processes. However in relation to bringing AI fashions and brokers, the trade took a extra cautious highway at the start. Now, extra monetary companies corporations, together with Brex, have launched AI-powered platforms and a number of agentic workflows.
Brex’s first foray into generative AI was with its Brex Assistant, launched in 2023, which helped clients automate sure finance and expense duties. It gives ideas to finish bills, routinely fills in info, and follows up on bills that violate insurance policies.
Reggio acknowledges that Brex Assistant works, however it’s not sufficient. “I feel to a point, it stays a little bit of a know-how the place we don't fully know the bounds of it," he mentioned. "There's fairly a lot of patterns that have to exist round it which are form of being developed by the trade because the know-how matures and as extra corporations construct with it."
Brex Assistant makes use of a number of fashions, together with Anthropic’s Claude and customized Brex-models, in addition to OpenAI’s API. The assistant automates some duties however continues to be restricted in how low-touch it may be.
Reggio mentioned Brex Assistant nonetheless performs an enormous position within the firm’s autonomy journey, primarily as a result of its Agent Mesh product flows into the applying.
Agent Mesh to exchange orchestration
The consensus within the trade is that multi-agent ecosystems, during which brokers talk to perform duties, require an orchestration framework to information them.
Reggio, however, has a unique take. "Deterministic orchestration infrastructure … was an answer for the issues that we noticed two years in the past, which was that brokers, identical to the fashions, hallucinate lots,” Reggio mentioned. “They're not excellent with a number of instruments, so it’s good to give them these levels of freedom, however in a extra structured, inflexible system. However because the fashions get higher, I feel it's beginning to maintain again the vary of potentialities which are increasing.”
Extra conventional agent orchestration architectures both deal with a single agent that does every part or, extra generally, coordinator/orchestrator plus instrument brokers that explicitly outline workflows. Reggio mentioned each frameworks are too inflexible and resolve points extra generally seen in conventional software program than in AI.
The distinction, Reggio argues, is structural:
-
Conventional orchestration: predefined workflows, central coordinator, deterministic paths
-
Agent Mesh: event-driven, role-specialized brokers, message-based coordination
Agent Mesh depends on stitching collectively networks of many small brokers, every specializing in a single process. The brokers, as soon as once more utilizing the hybrid mixture of fashions as with the Brex Assistant, talk with different brokers “in plain English” over a shared message stream. A routing mannequin rapidly determines which instruments to invoke, he mentioned.
A single reimbursement request triggers a number of duties: a compliance verify to align with expense insurance policies, price range validation, receipt matching, after which cost initiation. Whereas an agent can definitely be coded to do all of that, this technique is “brittle and error-prone,” and it responds to new info shared by means of a message stream anyway.
Reggio mentioned the concept is to disambiguate all of these separate duties and assign them to smaller brokers as a substitute. He likened the structure to a Wi-Fi mesh, the place no single node controls the system — reliability emerges from many small, overlapping contributors.
“We mainly discovered a extremely good match with the concept of embodying particular roles as brokers on high of the perfect platform to handle particular duties, very like the way you may delegate accounts payable to at least one crew versus expense administration to a different crew,” Reggio mentioned.
Brex defines three core concepts within the Agent Mesh structure:
-
Config, the place definitions of the agent, mannequin, instruments and subscription stay
-
MessageStream, a log of each message, instrument name and state transition
-
Clock, which ensures deterministic ordering
Brex additionally constructed evaluations into the system, during which the LLM acts as a decide, and an audit agent evaluations every agent’s selections to make sure they adhere to accuracy and behavioral insurance policies.
Success thus far
Brex says it has seen substantial effectivity beneficial properties amongst its clients in its AI ecosystem. Brex didn’t present third-party benchmarks or customer-specific information to validate these beneficial properties.
However Reggio mentioned enterprise clients utilizing Brex Assistant and the corporate’s machine studying techniques “are capable of obtain 99% automation, particularly for patrons that basically leaned into AI.”
It is a marked enchancment from the 60 to 70% Brex clients who have been capable of automate their expense processes earlier than the launch of Brex Assistant.
The corporate continues to be early in its autonomy journey, Reggio mentioned. But when the Agent Mesh method works, probably the most profitable end result could also be invisible: staff not desirous about bills in any respect.
