Post by DataDome

14,984 followers

Write a prompt, walk away, and let an LLM get to work. That's all it takes to automate a refund fraud attempt today, all while keeping the product in the process. The agent hammers the chatbot, retries when a session closes, and keeps going until it gets the answer it wants. Most organizations have no way of knowing whether a chat session is being driven by a real customer or an AI agent sent to act on their behalf. And if it is an agent, they now have to determine its intent. šŸ”— Check out Simon Marchand, CFE & Jordan Harris' full conversation, linked in the comments.

Post content

Video Content