Brex PM interview: the business model is the frame for every answer

Brex PM interviews are not hard in a puzzle sense. They are hard in a domain-knowledge sense. The bar is whether you can walk into a product sense or business sense conversation and immediately frame your answer around Brex’s actual revenue model: float income on card spend volume, software subscription fees on Brex Empower, and credit risk managed against a portfolio of high-growth startups and mid-market companies. Every strong answer connects a product feature to one of three levers: grow card spend, reduce credit risk, or expand Empower seats. Every weak answer imports a consumer product mental model into a B2B fintech context and gets called on it.

The process runs recruiter screen to product sense call to virtual onsite. PM roles are rated 3.1-3.3 out of 5 for difficulty, and 80% of candidates report a positive experience. That positive experience rating is real: the interviewers are practitioner PMs, not riddle-setters.

The interview stages

Recruiter screen (30 min). Background and motivation check. Brex recruiters will ask why you want to work on B2B financial infrastructure rather than consumer. “I like fintech” is not enough. Know what Brex Empower does (AI-native spend management: expense automation, policy enforcement, bill pay, travel, treasury) and be able to name at least one product area where you would want to build.

Product sense round (45-60 min live call). A PM from a different team runs this. Per recruiter guidance, the question is Brex-specific, not a generic “improve Google Maps” prompt. Expect a prompt drawn from corporate card controls, expense policy enforcement, reimbursements, bill pay, travel booking, or treasury workflows. This is not a take-home. Do not prepare a multi-day document exercise. Prepare to run a live structured design conversation.

One important correction that circulates in prep communities: questions about designing a personal finance app for college students are outdated. Brex sold its consumer business to Empower Finance in 2020. Your entire product sense frame should live in B2B.

Virtual onsite. Two primary rounds:

Execution round with an engineering director. Root cause analysis, metrics definition, prioritization under resource constraints. This is where fintech domain knowledge gaps get exposed.
Business sense round with a PM lead from a different team. Can you model the unit economics of a spend management feature and name the right guardrail metrics? This is Brex’s version of a strategy round. It is explicitly connected to the business model, not abstract strategic frameworks.

The business sense round: what it is actually testing

“Business sense” at Brex means: do you understand that the corporate card generates revenue through interchange (a percentage of every transaction), and that Empower generates revenue through per-seat software fees? A product decision that increases card spend volume or reduces fraud and chargebacks has a clear revenue connection. A product decision that only improves employee satisfaction scores does not, unless you can trace that satisfaction to retention of seats or increased spending.

strong

"The question is how to measure impact of a new expense policy enforcement feature in Brex Empower. I'd start by grounding in the business model. Policy enforcement reduces two things: fraud risk (which protects Brex's balance sheet on the card side) and manual finance team work (which drives Empower retention and expansion). So I need metrics that capture both. Primary metric: percentage of out-of-policy transactions automatically blocked or rerouted without manager intervention. This measures automation quality, not just coverage. Brex's own data shows AI already handles roughly 70% of expenses with managers reviewing 6x faster. A regression on that 6x number signals the feature is creating friction rather than removing it. Secondary metric: time-to-resolution for flagged expenses. Guardrail: 30-day delinquency rate on card accounts where the enforcement change was applied. The feature cannot shift credit risk onto Brex's balance sheet by over-blocking legitimate spend and disrupting cash flow. Business outcome: Empower seat retention at the 6-month mark for accounts where the feature activated. This connects the feature to all three revenue levers. I'm not measuring DAU or NPS because those are vanity metrics in a tool that employees open only when they have an expense to submit."

weak

"I'd track daily active users, employee satisfaction scores, and whether expense submission time goes down." This fails on every axis. DAU is a vanity metric for a B2B workflow tool: employees don't open Brex Empower for fun, they open it when they have an expense. Satisfaction scores are lagging and don't connect to the revenue model. Submission time going down could mean employees are bypassing policy enforcement, which is the opposite of success. This answer comes from someone who applied a consumer product framework to an enterprise fintech context. The interviewer is listening for whether you understand float income and software seat economics before you name a single metric.

Fintech domain knowledge: what you need

You do not need a background in financial services. About an hour of reading closes most gaps. The concepts that matter for Brex interviews:

Double-entry accounting: every transaction creates a debit and a credit entry. Reconciliation is the process of making them match. This is core to what Brex Empower’s accounting close feature does (3x faster close in 2025 with AI).
Authorization vs. capture: a card swipe authorizes a hold; capture is when the merchant actually collects the funds. Policy enforcement can happen at authorization, before any money moves.
Interchange and float: Brex earns interchange on every card transaction (a percentage of spend). Float income comes from holding customer deposits before disbursing. Both are sensitive to spend volume.
Chargebacks and fraud: disputes that reverse a completed transaction. Fraud detection that prevents chargebacks protects revenue and limits credit risk.
Idempotency: a payment submitted twice should only execute once. This is table stakes for any PM working on bill pay or reimbursements.

The 2026 context: Capital One acquisition and the Empower AI platform

Capital One acquired Brex in a $5.15 billion cash-and-stock deal that closed April 7, 2026. Brex retains product and startup strategy autonomy in the near term, but the acquisition creates a tension that sharp candidates should be able to articulate.

Brex was built on an insight: high-growth startups are better credit risks than traditional underwriting models suggest, because their trajectory matters more than their history. Capital One operates at commercial banking scale with a more conservative risk posture. In every product sense or business sense question you answer in 2026, you may encounter a version of this tension: where does Brex’s startup-first credit philosophy hold firm, and where does it adopt more conservative guardrails at enterprise scale? Candidates who can name this tension and reason through it will stand out. Candidates who ignore the acquisition will not.

On the product side, Brex Empower’s AI agent platform is the strategic core. The platform freed $163 million in annual salary costs for customers and saved 208,000 hours per month as of 2025. Roughly 70% of expenses are handled by automation. The implication for PM candidates: “lovable” in this context has nothing to do with delightful consumer UI. The primary actor in many Empower workflows is an AI agent submitting, categorizing, and routing expenses on behalf of a human. Lovable means the system anticipates the policy edge case before the employee hits it, routes the exception without a manager touching it, and closes the audit trail automatically. Human review is a fallback, not the design target.

Brex powers OpenAI’s global spend and financial operations, and its partnership with Fifth Third Bank adds $5.6 billion in commercial card volume. The PM interview reflects the sophistication of the customers Brex serves: the world’s fastest-scaling technical companies, many of which are also operating AI agents that submit expenses autonomously. A PM who walks in without a view on agentic finance workflows is behind.

What to prepare before the product sense call

Go in with a view on one or two product areas where you would want to build at Brex. The strongest candidates have done one thing: they used Brex (or read detailed customer reviews) and found a real gap in the expense submission, reimbursement, or treasury workflow. The product Brex is interviewing for in 2026 is an AI-native B2B spend management platform inside a commercial banking context, and the interview reflects that.

See fintech PM interviews, feasibility is free, and lovable, not just usable for the frameworks that underlie what Brex is looking for.