Apate's AI agents pose as scam victims to waste fraudsters' time and harvest intel

Australian startup Apate.ai has turned scam-baiting into a commercial product: according to TechCrunch, the company runs thousands of AI agents whose job is to be targeted by fraudsters, keep them talking, and pull intelligence from the exchange that banks and telecom operators can act on.

Decoys at scale

Apate was founded in 2024 on the back of research at Macquarie University in Sydney. Its agents impersonate ordinary people — varying age, gender, accent and language — and engage scammers across channels: they answer fraud calls, reply to shady texts and emails, follow phishing links and join social media groups where scammers hunt for targets.

The company says its fleet has burned the equivalent of 17 years of scammers' time across roughly 2.5 million conversations. The bigger payoff, though, is information. The agents try to get scammers to reveal details that normally surface only after a real victim has been defrauded — bank accounts, crypto wallet addresses, phone numbers, malicious websites and the brands being impersonated.

Customers, which include banks, telecom operators, government agencies and utilities, receive that data through an insights portal offering analytics on scams aimed at their users, employees and brands. They can also assign agents specific missions, such as obtaining an account number or a wallet address. Text agents can sweep forums, investment groups and social platforms for signs of scam activity, then engage under a fabricated identity to surface details about who is running it.

Built from a 44-minute phone call

The idea came from an actual scam call. According to TechCrunch, founder and CEO Dali Kaafar took a call from a scammer during a family picnic and played along for 44 minutes to amuse his children. Afterwards he registered two things: those were 44 minutes the scammer had not spent on a real victim, and 44 minutes of his own time gone. That led to research at Macquarie using machine learning and natural language processing to analyse scam calls and improve detection and prevention.

Co-founder and chief commercial officer Brad Joffe frames the company as pre-emptive rather than reactive. Telcos block numbers and scammers simply acquire new ones; banks watch transactions only after a fraudster has already reached the customer. Apate's agents instead engage scammers directly and gather information before anyone loses money, reportedly talking to tens of thousands of scammers every day across calls, messages, email and the web. The company powers them with a mix of frontier models and its own fine-tuned models, trained on its agents' recorded conversations with scammers plus the Macquarie research.

Reported results and funding

Per TechCrunch, named customers include Commonwealth Bank of Australia and TPG Telecom. TPG has diverted nearly 590,000 calls to Apate's agents, Commonwealth reports a 76% drop in publicly reported scam losses, and an unnamed Australian bank used Apate's outbound agents to disrupt 22 scam campaigns. These figures come from the company itself and have not been independently verified.

Apate raised $8.15 million in seed funding in August from Lobby Capital, OIF Ventures, Investible, Concept Ventures and Baobab Ventures, bringing total funding to just under $10 million. The roughly 20-person company plans to spend the round on product development and international expansion, including a London office and a larger US presence. It was also selected as a finalist in the Startup Battlefield 200 at TechCrunch Disrupt. The name references the Greek goddess of deception, and the company tracks scammers swearing at its bots as a measure of frustration — even celebrating each expletive as a success signal.

Why it matters

AI has sharply lowered the cost of running fraud at scale. TechCrunch cites FBI figures of more than 22,000 complaints of AI-related crimes in 2025, causing about $893 million in losses, with over $632 million of that from investment scams where fraudsters use AI to sustain conversations and deepfake the voices and appearances of trusted figures. Apate applies the same capability in reverse: if models can convincingly impersonate celebrities and executives, they can also impersonate victims, making decoys cheap and effectively unlimited. For anti-fraud teams stuck blocking numbers and reviewing transactions after the fact, intelligence gathered from inside an active scam is a fundamentally different input — and if the company's numbers hold up, it turns a long-standing internet hobby into a scalable commercial defence.