Apate.ai, an Australian startup founded in 2024 from research at Macquarie University, has developed a unique approach to combatting fraud by deploying thousands of AI agents that pose as potential scam victims. These AI agents interact with scammers through calls, messages, emails, and social media groups, keeping scammers engaged and tricking them into revealing crucial information usually only accessed after a successful scam. By impersonating a wide range of profiles across different languages and accents, Apate’s fleet has wasted the equivalent of 17 years of scammers’ time in over 2.5 million conversations.

The data gathered by Apate’s AI agents includes sensitive details such as bank accounts, crypto wallet addresses, malicious websites, impersonated brands, and other insights into scam tactics. This intelligence is then shared with Apate’s clients, which include major banks, telecoms, utilities, and government agencies, enabling them to strengthen their fraud detection and prevention strategies. Notable customers like Commonwealth Bank of Australia and TPG Telecom have reported significant positive results, including a reported 76% reduction in scam-related losses at Commonwealth Bank.

The concept arose from an experience by Apate’s founder Dali Kaafar, who once amused his family by extending a scam call for 44 minutes to waste the scammer’s time. This led to the idea of automating the process to both occupy scammers and extract valuable intelligence. Apate’s technology employs both state-of-the-art AI models and tailored algorithms developed from Kaafar’s academic research, allowing agents to conduct extensive, multilingual conversations and even undertake customized missions such as securing bank or wallet numbers from fraudsters.

Apate recently raised $8.15 million in seed funding to expand its operations internationally, with ambitions to grow in the U.S. and open a London office. The company’s co-founder and chief commercial officer, Brad Joffe, emphasized that Apate goes to the scammer first, in contrast to industry methods that typically intervene only after fraud has occurred. The startup, a finalist in the 2026 TechCrunch Disrupt Startup Battlefield 200, humorously celebrates when scammers curse their AI agents, a testament to its effectiveness in frustrating criminal activity.