Apate sends AI agents undercover to get scammed so they can gather fraud intelligence
While the vast majority of the technology sector is racing to deploy AI agents to streamline supply chains, accelerate software development, or automate mundane office workflows, one Australian startup is taking a decidedly different approach. Apate.ai is weaponizing artificial intelligence by sending its agents undercover to intentionally get scammed.
Founded in 2024, the company emerged from foundational research conducted at Macquarie University in Sydney. The startup deploys a massive fleet of thousands of AI agents designed to masquerade as vulnerable, unsuspecting targets. These digital decoys are programmed to engage with the criminal underworld by answering unsolicited scam calls, replying to phishing emails, clicking on malicious links, and infiltrating private social media groups used by fraudsters to coordinate their activities.
Turning the Tables on Fraudsters
The agents are sophisticated, capable of adopting diverse personas—varying in age, gender, accent, and language—to keep scammers engaged for as long as possible. The goal is simple: waste the criminal's time while extracting actionable intelligence. To date, Apate’s fleet has successfully tied up scammers for a cumulative 17 years across 2.5 million individual conversations.
For Apate’s primary clientele—which includes major banks, telecommunications providers, government agencies, and utility companies—this is not just about wasting time. It is about data. By keeping fraudsters on the line, the agents coax them into revealing critical infrastructure details, such as:
- Financial identifiers: Bank account numbers and cryptocurrency wallet addresses.
- Technical infrastructure: Malicious URLs, phishing domains, and burner phone numbers.
- Tactical intelligence: The specific brands being impersonated and the methodologies behind current scam campaigns.
This intelligence is funneled directly to Apate’s customers, allowing them to proactively update their anti-fraud defenses before their own users fall victim to these schemes.
A Response to the AI Crime Wave
The necessity for such a tool has never been greater. As AI lowers the barrier to entry for cybercriminals, the financial stakes have skyrocketed. According to FBI data, 2025 saw over 22,000 reports of AI-related criminal activity, resulting in approximately $893 million in losses. A significant portion of this—$632 million—was attributed to investment scams where fraudsters utilized AI to conduct convincing, long-form conversations and deploy deepfake audio or video of trusted public figures.
"Most of the industry works after the fact. Telcos block numbers, but scammers just get new ones. Banks watch transactions, but by then the scammer has already reached the customer. We go to the scammer directly and get the information before anyone gets hurt." — Brad Joffe, Co-founder and CCO of Apate.
The Origin Story: A 44-Minute Lesson
The genesis of Apate lies in a personal experience. Several years ago, CEO and founder Dali Kaafar received a scam call during a family picnic. Rather than hanging up, he decided to play along for the amusement of his children, keeping the scammer on the line for 44 minutes.
While the experience provided a laugh, it also provided a revelation: those 44 minutes were time the scammer could not spend victimizing someone else. Kaafar began to wonder if this "scam-baiting" could be automated and scaled to provide real-world security benefits. This curiosity led to his academic research at Macquarie University, where he utilized machine learning and natural language processing to decode the mechanics of scam calls.
Operational Intelligence and Impact
Apate utilizes a hybrid approach, combining frontier AI models with its own proprietary, fine-tuned systems. The company continuously trains its agents on the data harvested from millions of interactions with real-world scammers.
The results are already tangible. TPG Telecom has reportedly diverted nearly 590,000 calls to Apate’s agents, while the Commonwealth Bank of Australia has observed a 76% reduction in publicly reported scam losses. Furthermore, an unnamed Australian financial institution has successfully utilized Apate’s outbound agents to disrupt 22 active scam campaigns.
Key Performance Metrics
- Total Time Wasted: 17 years of scammer time.
- Engagement Volume: 2.5 million conversations.
- Strategic Impact: 76% drop in reported losses for key banking partners.
Scaling the Deception
Investors are clearly convinced that there is a viable business model in outsmarting the bad guys. In August, Apate secured $8.15 million in seed funding from a group including Lobby Capital, OIF Ventures, Investible, Concept Ventures, and Baobab Ventures. This brings the 20-person company’s total funding to just under $10 million.
The startup plans to use this capital to fuel international expansion, with immediate goals including the opening of a London office and a significantly larger footprint in the United States.
Named after the Greek goddess of deception, Apate embraces the irony of its mission. The team even keeps a tally of the profanities directed at their bots by frustrated scammers. For the team at Apate, every swear word is a badge of honor—a metric of success that proves their agents are doing exactly what they were built to do: getting under the skin of those who seek to exploit others.