AI Agents That Get Scammed on Purpose: Apate Has Wasted 17 Years of Scammers Time Across 2.5 Million Conversations

Most AI startups build agents to work for you. Apate, a Sydney-based startup, builds agents to work against scammers — by getting scammed. The company, founded in 2024 on research from Macquarie University, deploys thousands of AI agents that masquerade as potential victims, answering scam calls, replying to suspicious messages and emails, following phishing links, and even joining social media groups to keep fraudsters busy.

The agents play people of all stripes: different ages, genders, accents, and languages, holding personas consistent over long, multi-turn conversations. The goal is not to lose money but to make scammers expose themselves — bank accounts, crypto wallet addresses, phone numbers, malicious websites, the brands they impersonate, and the software they ask victims to install. In a single victim's report these details are fragments; across thousands of undercover agents they become a structured intelligence graph.

The company says the results are measurable in an unusual unit: time. Apate claims its fleet has now engaged in 2.5 million conversations and wasted 17 years of scammers' time — company-reported figures, not independently audited. Co-founder and chief commercial officer Brad Joffe says the agents talk to tens of thousands of scammers every day across calls, messages, email, and the web.

Apate's customers are banks, telecom operators, government agencies, and utilities, and the company points to concrete outcomes it attributes to its fleet: TPG Telecom has diverted nearly 590,000 calls to Apate agents, Commonwealth Bank of Australia has seen a 76% drop in publicly reported scam losses, and an unnamed Australian bank used outbound agents to disrupt 22 scam campaigns. All of those figures come from the company.

The idea started with a family picnic. Founder and CEO Dali Kaafar, realizing an unknown caller was a scammer, decided to play dumb for his kids and kept the scammer talking for 44 minutes. Those were 44 minutes the fraudster did not spend on a real victim — and 44 minutes of Kaafar's own wasted time, which raised a question: could this be automated, and could it produce something useful? The answer became a research program at Macquarie University, and then a company.

Technically, Apate runs a mix of frontier models and its own fine-tuned models, trained on the conversations its agents have had with scammers plus the underlying Macquarie research. Customers get an insights portal with intelligence and analytics on scams targeting their users, employees, or brands, and can even assign agents specific missions, such as obtaining a bank account number or a crypto wallet address.

Investors are betting the deception business pays. 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 for the 20-person company. The money is earmarked for product development and international expansion, including a London office and a larger US presence. Apate is also a finalist in this year's Startup Battlefield 200 at TechCrunch Disrupt.

The timing is not accidental. The FBI received more than 22,000 complaints of AI-related crimes in 2025, with roughly $893 million in losses, over $632 million of which came from investment scams where fraudsters use AI to run conversations and deepfake voices and faces. Defenses that react after the fact — banks blocking transactions, telcos blocking numbers — iterate in days while scam operations iterate in hours. Apate's pitch is to flip the direction of travel: instead of waiting for victims, send the probe to the scammer first. If generative AI has industrialized fraud, the countermeasure, it turns out, looks surprisingly like the same technology pointed the other way.