← All implementations

Brex · Platform

Internal Agent Platform

Brex built an internal platform in Retool where operations staff build, test and deploy agents for KYC checks, card disputes, support quality assurance and collections. A roughly 25-person systems engineering team maintains it, and analysts still review every case the agents do not handle with high confidence.

1

Supporting infrastructure

This entry describes supporting infrastructure that other work builds on. The catalog classifies it as a platform. The record also reports a workflow.

Approach type
Platform
Work
Finance ops, Support, Customer success
Human involvement
Human in loop
Invocation
Interactive
Interfaces
Slack, Internal ui
Deployment stage
Scaled
Evidence strength
Secondary only
Entry reviewed

How it works

The workflow the sources report for this implementation.

Non-technical ops staff design prompts, test across models, deploy with QA oversight

1

Product features exposed to internal agents through one MCP server

1

Where people stay involved

  • internal operations request → completed operationContinuous steering · Level 2

Level 2 for internal operations request → completed operation; human attention boundary: continuous-steering.

1

Observation date
2025-09-25

Implementation details

Sandbox

Retool-hosted runtime (no bespoke execution env described)

1

Harness

Retool-based builder with prompt management and multi-model testing/evaluation; built by a ~25-person systems-engineering team

1

Model

Multi-model

1

Interfaces

slack, internal-ui

1

Tool access

An MCP server exposes external product features to the internal platform; new product tools become internally available immediately; invoked via Slack /c1

1

Knowledge

Standard operating procedures uploaded as a knowledge base, such as 100-page dispute guides; customer account data

1

Credentials

SSO via internal Retool proxies (no per-user accounts); ConductorOne access management; Okta auth; data classified by risk; ≤30-day retention, no training on inputs

1

Reported results and limitations

The catalog records what the sources report, with the scope and the denominator of every figure. A qualification below limits the figure it sits under.

Reported metrics

Headline claim

Dispute processing time fell from three hours to three seconds

1

The source does not report the denominator of this figure.

Reported by
Brex
Scope
Dispute-submission preparation using the internal agent platform; not end-to-end chargeback resolution
Observation date
2025
Key observation

50%+ of customer-support cases resolved by chatbot as first touch

1

Reported by
Brex
Scope
Customer-support cases resolved by the chatbot at first touch
Denominator
Customer-support cases; exact sample size not provided
Observation date
2025
Key observation

Dispute processing: 3 hours → 3 seconds

1

The source does not report the denominator of this figure.

Reported by
Brex
Scope
Dispute-submission preparation using the internal agent platform; not end-to-end chargeback resolution
Observation date
2025
Key observation

QA covers every support interaction; one person using AI instead of five QA specialists

1

Reported by
Brex
Scope
Quality assurance of customer-support interactions
Denominator
Every support interaction
Method
Agent applies the quality rubric to every response; one person oversees instead of five QA specialists
Observation date
2025
Key observation

KYC adverse-media accuracy 85% → 88%

1

The source does not report the denominator of this figure.

Reported by
Brex
Scope
KYC adverse-media classification; human versus agent accuracy
Observation date
2025

Lessons and interpretation

Target 40% automation, not 100%; the last mile drives investments that often yield zero value

1

Brex reports that it manages the internal platform with practices used for an external product

1

Don't skip the human in the middle; end-to-end automation fails on accuracy

1

Map workflows to the discrete steps a human would take, then translate each to LLM instructions

1

Architecture beats vendor; legal/data approvals should follow data-handling characteristics, not the tool name

1

Sources and research details

Citations link to the original publisher. Each source also keeps a preserved copy in the repository, so a changed or removed page stays checkable.

  1. Agent, Human, Ops: How Brex Is Changing Roles and Workflowshttps://www.firstround.com/ai/brexCase study · Independent secondary · Last source verification: 2026-08-31
    Preserved copy ↗
Research details for every claim on this page
  1. Summary
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  2. Headline claim
    Statement type
    Metric
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
    Reported by
    Brex
    Scope
    Dispute-submission preparation using the internal agent platform; not end-to-end chargeback resolution
    Denominator
    Not reported
    Method
    Not reported
    Observation date
    2025
  3. Sandbox
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  4. Harness
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  5. Model
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  6. Interfaces
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  7. Tool access
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  8. Knowledge
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  9. Credentials
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  10. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  11. Supporting component
    Statement type
    Fact
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
  12. Key observation
    Statement type
    Metric
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
    Reported by
    Brex
    Scope
    Customer-support cases resolved by the chatbot at first touch
    Denominator
    Customer-support cases; exact sample size not provided
    Method
    Not reported
    Observation date
    2025
  13. Key observation
    Statement type
    Metric
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
    Reported by
    Brex
    Scope
    Dispute-submission preparation using the internal agent platform; not end-to-end chargeback resolution
    Denominator
    Not reported
    Method
    Not reported
    Observation date
    2025
  14. Key observation
    Statement type
    Metric
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
    Reported by
    Brex
    Scope
    Quality assurance of customer-support interactions
    Denominator
    Every support interaction
    Method
    Agent applies the quality rubric to every response; one person oversees instead of five QA specialists
    Observation date
    2025
  15. Key observation
    Statement type
    Metric
    Provenance
    Reported
    Confidence
    Medium
    Confidence reason
    A linked participant or independent source reports the claim.
    Reported by
    Brex
    Scope
    KYC adverse-media classification; human versus agent accuracy
    Denominator
    Not reported
    Method
    Not reported
    Observation date
    2025
  16. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  17. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  18. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  19. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  20. Lesson
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The catalog derives this observation from the linked sources.
  21. Operating model assessment
    Statement type
    Inference
    Provenance
    Catalog judgment
    Confidence
    Medium
    Confidence reason
    The secondary report describes a human remaining in the middle of the workflow, but does not fully specify each review surface.
    Observation date
    2025-09-25

The catalog records no related implementation for this entry yet.

Back to the directory →