Interview Questions › Root Cause Analysis

Rising fraud, higher review load, and declining trust simultaneously

Your trust & safety team faces a three-front crisis: fake host listing fraud is up 40%, manual agent review backlog surged by 300% causing 48h host verification delays, and guest post-checkout trust ratings dropped to a 3-year low. Dissect this systemic breakdown.

Context

Scammers are using AI-generated photos and virtual phone numbers to bypass initial signup verification.

Clarifying questions to ask first

• How are fraudulent hosts cashing out before guests discover the fake property?
• Which verification gate has the highest false-positive rate?

How to approach it

1. Map the Host Onboarding & Payout Security Funnel: Account Creation -> ID Verification -> Listing Address Proof -> Photo Upload -> Bank Payout Verification -> First Booking Escrow
2. Identify vulnerability exploits: Virtual numbers, synthetic identities, unverified GPS addresses, fake review rings
3. Analyze why automated risk scoring failed, flooding manual human review teams
4. Architect an end-to-end trust overhaul balancing friction for legitimate hosts vs fraud prevention

Framework: Fraud Vector Vulnerability Mapping -> Risk Engine Thresholds vs Human Ops Capacity -> Payout Escrow Defense -> Holistic Trust Architecture

Sample answer

• Immediate payout release upon booking confirmation allowed scammers to drain funds before guest check-in
• AI image generation tools defeating duplicate image hash detection algorithms
• Risk model flagging too many legitimate new hosts, causing ops queue overload

1. Vulnerability isolated: Payouts were released 24h after booking rather than 24h after physical guest check-in; fraud syndicates booked their own fake listings with stolen credit cards and extracted instant cash

Critical payout policy loophole identified; enforced payout release only 24h post-physical guest check-in, integrated automated utility bill address verification, and deployed reverse AI image forensics.

What interviewers look for

• Fraud Chargeback Loss Rate
• Manual Review Backlog (Hours)
• Time-to-Publish Host Listing
• Guest Safety NPS