An Australian technology firm has developed a network of nearly 200,000 AI-driven personas designed to engage with online fraudsters, wasting millions of hours of scammers' time each month while extracting actionable intelligence for financial institutions and telecommunications providers.
Apate, founded in 2023, operates a global system of AI characters that simulate gullible victims in phone conversations and social media interactions. The company's monthly performance metrics include tracking the frequency with which frustrated scammers resort to profanity during failed attempts to deceive the bots. "We're likely the only company in the world that includes the number of F-words scammers direct at our bots as a KPI," said Apate founder Dali Kaafar.
During a six-week period ending December 2025, Apate's bots conducted approximately 600,000 scam calls for a single Australian telecom provider, TPG. The engagements consumed an estimated 500 days of scammers' time, equating to roughly $13 million in potential savings by preventing fraudulent activity, according to company estimates. The AI systems operate across multiple channels, including WhatsApp and Telegram, where they function as honeypots to lure and monitor fraudulent operations.
The initiative originated from Kaafar's personal experience in 2021, when he received a scam call while on a family outing. After engaging the fraudster for 44 minutes, he recognized the potential for scalable technological intervention. "As I hung up, I thought: if I could do this for fun, imagine what technology could achieve at scale," Kaafar said. The concept was developed with doctoral students at Macquarie University, leading to initial funding from the Office of National Intelligence and subsequent commercialization.
Apate now collaborates with major banks in Australia, the UK, South Africa, and Southeast Asia. The company's AI models, trained on hundreds of hours of recorded scam baiting sessions, employ counter-strategies to maintain realistic interactions. "The bots sound exactly like you and I, with identifiable vocal tics and accents," Kaafar noted. Their primary objectives include disrupting scam operations and gathering intelligence on fraudster tactics, wallet addresses, and operational networks.
In July, Apate's systems identified a marketplace in India where brokers solicited verified bank accounts for scam proceeds, offering commissions of up to 5% paid in USDT. The company's research suggests that 20% to 30% of scam text conversations already involve AI, though Kaafar maintains that defensive AI systems hold an advantage in this emerging arms race. "Game theory suggests that AI defenders extracting intelligence have an inherent edge over AI scammers attempting to elicit actions," he said.
The broader implications extend beyond individual engagements, as scamming represents a $1.24 trillion global industry. Apate's approach aligns with growing efforts by financial institutions and regulators to combat increasingly sophisticated fraud networks.
