Hacking AI customer service agents

Intigriti's blog post walks through real attack techniques against AI customer service agents, covering prompt injection, data leakage, and privilege escalation. It's a practical security checklist for anyone deploying LLM-based customer-facing systems.