Most companies use AI agents to ship products faster or speed up workflows. Apate.ai infiltrates them to get scammed.
The Australian startup, founded in 2024 based on research conducted at Macquarie University in Sydney, has deployed thousands of AI agents posing as innocent victims. Agents can answer scam calls, respond to suspicious messages and emails, follow phishing links, and join social media groups to lure scammers. They pretend to be different people of different ages, genders, and accents, speak different languages, and try to keep the scammer around longer in order to learn something useful.
The company says its agents have wasted 17 years of scammers’ time in 2.5 million conversations. More importantly for Apate’s customer base of banks, carriers, government agencies, and utilities, the undercover agents are forcing the fraudsters to reveal things that victims typically only learn after they’ve been scammed: bank accounts, crypto wallets, phone numbers, malicious websites, spoofed brands, and other fraudulent details.
Apate, one of this year’s TechCrunch Disrupt Startup Battlefield 200 finalists, provides that information to its customers, who use it to inform their anti-fraud efforts. (The company will present a panel of VC judges at Disrupt, which runs through Thursday, October 15th, at Moscone West in San Francisco this week.)
This is an interesting evolution in fraud, but not all that surprising considering how much power AI has given scammers and hackers. The FBI received more than 22,000 charges related to AI-related crimes in 2025, resulting in approximately $893 million in losses. More than $632 million of that was in investment fraud, with fraudsters using AI to drive conversations and deepfaking the voices and appearances of celebrities, executives, and other trusted figures, according to the FBI.
For a long time, people have made it a hobby to intentionally attract scammers. Sometimes it’s hilarious, like having a scammer read an entire book out loud or get an outrageous tattoo. Now, with AI becoming more and more capable, the tables are turning.
Oddly enough, the idea behind Apate came from an actual scam call.
A few years ago, Dali Kaafar, founder and CEO of Apate, was having a picnic with his family when he received a call from an unknown number. Realizing that the person on the other end of the phone was a scammer, he decided to play dumb for the benefit of his children and managed to keep the call going for 44 minutes.
Everyone laughed, but Kaafal realized that that was 44 minutes that the scammer didn’t spend before calling the real victim. It was also 44 minutes of wasted time, he thought. “Can we automate this and learn something useful in the process?”
The idea eventually led to research work at Macquarie University, where Karfer and colleagues used machine learning and natural language processing to uncover the content of fraudulent calls and develop better ways to detect and prevent fraud.
“Most of the industry is reactive. The carrier blocks the number, but the fraudster just gets a new number. The bank monitors the transaction, but by then the fraudster has already contacted the customer. We go to the fraudster face-to-face and get the information before anyone is harmed,” co-founder and chief commercial officer Brad Joffe told TechCrunch. “Our AI agents talk to tens of thousands of scammers every day via phone, text, email, and the web. What we give our customers is what they can get out of those conversations and what they can do with them.”
Joffe said Apate uses both frontier models and proprietary fine-tuned models to power its agents and analytics. The company trains its models based on conversations agents have with fraudsters and research Kafer conducted at Macquarie.
In addition to access to Apate’s fleet of agents, customers also have access to an insights portal that provides fraud intelligence and analysis targeting end users, employees, brands, products, and more. Customers can also assign specific missions to these agents, such as retrieving bank account numbers or cryptocurrency wallet addresses. Its text agents can search forums, investment groups, and social platforms for potential scams and use a forged identity to join the conversation and reveal details about the scammer.
The startup’s customers include the Commonwealth Bank of Australia, TPG Telecom and other utilities, telcos and government agencies. Mr Joffe said TPG had routed about 590,000 calls to Apate’s agents, resulting in a 76% reduction in publicly reported fraud losses in the Commonwealth, and that an unnamed Australian bank with which the startup does business had used outbound agents to thwart 22 fraud campaigns.
Investors are betting on scammers to make money. Apate raised $8.15 million in seed funding in August from Lobby Capital, OIF Ventures, Investible, Concept Ventures, and Baobab Ventures. Joffe said the 20-employee company has raised just under $10 million in total and plans to spend the latest round on product and international expansion, including opening an office in London and building a larger presence in the United States.
“Apate is named after the Greek goddess of deception, which felt appropriate given the activity of our agents,” Joffe said, adding that the startup also features scammers who taunt bots in their pitch materials.
“We track that as an expression of frustration, and at Apate every bad word is celebrated. We’re probably one of the only companies in the world that celebrates when people badmouth our products.”
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
