On April 1, 2026, Visa lowered the merchant threshold in its Acquirer Monitoring Program from 2.20 percent to 1.50 percent. The program, known as VAMP, replaced two older programs and now folds fraud reports and disputes into one count-based ratio. That ratio is a fraud and risk product manager's whole job in miniature. Screen too little and losses climb past the threshold. Screen too much and you block real buyers.
Interviews for these roles test whether you can hold both sides of that tradeoff at the same time. The questions rarely ask for a single right answer. They ask how you reason about money, customers, and rules under pressure. This guide covers what those interviews look at and how to prepare for each part.
What a fraud and risk PM owns
A fraud and risk PM sits between the money and the customer. On one side is loss: chargebacks, stolen cards, account takeovers, and refund abuse. On the other side is friction, since every check you add can slow down a legitimate purchase. Your product is the set of rules, models, and review steps that decide which payments the system clears and which it holds for review.
Most teams measure this work with a few numbers. Fraud rate is the share of payments that turn out to be fraud. Approval rate is the share of legitimate payments that reach approval. False positive rate is the share of legitimate payments the system declines in error. The art is moving one number without wrecking the other numbers.
The core tradeoff question
The most common prompt has a simple surface. Fraud went up last quarter, and the interviewer wants your next move. A complete answer starts by asking what kind of fraud. Card testing, where someone runs many small charges to check stolen numbers, calls for one response. Friendly fraud, where a real customer disputes a real purchase, points to a different fix.
From there, a full answer weighs the cost of each option. Tightening a rule cuts fraud but also declines more real orders. You can put a dollar figure on the change. If tightening saves 200,000 dollars in fraud but blocks 500,000 dollars in sales, then that rule costs more than the fraud.
A helpful move is to name your decision rule in plain terms. You might say you will accept a rule change only when the fraud saved beats the sales lost by a set margin. Then you can walk through how you would test the rule: run it in shadow mode, compare the declines it would have caused against the fraud it would have caught, and ship it only when the trade clears your bar. Interviewers read that as a sign of real experience with risk models.
System design for money movement
Some loops go deeper into how you would build a fraud system. A common version asks you to design the checks for a new checkout flow. Start with the layers. A rules engine catches known patterns in milliseconds. A machine learning model scores the rest by risk. A manual review queue handles the small slice that sits in the middle. A feedback path sends confirmed fraud back into the model as training data.
The interviewer will probe the tradeoffs. Real-time scoring adds latency, and every extra millisecond at checkout can lower conversion. Batch scoring is cheaper but lets some fraud settle before the system flags the charge. Talk about where you would draw the line and why that spot makes sense for the business.
Data is the other half of the answer. A fraud model is only as sharp as its labels, and those labels arrive late, since a chargeback can land 60 days after the sale. Point out that lag and explain how you would handle the delay, maybe by mixing early fraud signals with slower confirmed labels. That kind of detail shows real work with fraud data.
Where compliance comes in
Fraud work shares a border with compliance, so expect a question or two about rules. Know the basics of KYC, the process of confirming a customer's identity, and AML, the work of spotting money laundering. You do not need a lawyer's depth. You do need to show that you know when to pull in the risk and legal teams.
The VAMP change is a clean example of the overlap. When Visa lowered the ratio, every acquirer had to hold its merchants to a tighter limit. A PM on that team has to cut the fraud and dispute rate without choking off sales. That is a compliance deadline and a product problem in the same sprint.
The behavioral round
Risk teams live with hard calls, so the behavioral round digs into judgment. A frequent question asks about a time you shipped something that later caused harm. Pick a real story with a clear metric and a fix you shipped in the next release. Owning the miss carries more weight than a record with no scars.
You may also face a question about conflict. Fraud teams push for tighter controls while growth teams push for fewer blocks. Describe a time you sat between those two goals and found a middle path with room for both goals. Winning the argument matters less than keeping both the loss rate and the customer in view.
How to get ready
Start with the money. Learn how interchange, chargeback fees, and fraud losses hit a payments P&L, because every risk decision lands on that statement. Read the public rules from the card networks, since a rule like VAMP will show up in real questions. Build one or two stories about a risk decision you made and the metric that moved as a result.
The thread through all of it is balance. A payments company needs to stop fraud and keep customers, and those two goals pull against each other every day. Show that you can hold the tension, put dollars on each side, and pick a point you can defend with numbers. These interviews are built to measure that skill.