Bunq Snitch

Slides

Bunq Snitches solves the gap between financial intention and real behavior. People set goals like spending less on fast food, impulse shopping less, or sticking to a budget, but those goals usually fail because there is no immediate accountability when a real transaction happens.

Our solution is an AI-powered accountability system connected to Bunq transaction events. Users create “snitches” with a goal and personality, and when a payment comes in, the snitch decides whether to stay quiet, warn the user privately, ask for an explanation, request proof, or escalate the behavior to a friend, partner, or group chat on Telegram.

AI is used as the reasoning layer. It interprets transactions in the context of a user’s goal, applies tone and escalation logic, and handles conversational follow-up when the user replies. Non-text modality is integrated through receipt images sent in Telegram: when a user says they can explain a spend, the agent asks for a photo of the receipt and uses that image as additional evidence in the decision flow.