Apate.ai Deploys AI Agents to Impersonate Victims and Gather Scam Intelligence
Australian startup Apate.ai has deployed thousands to approximately 350,000 AI agents that impersonate ordinary victims of various ages, accents, and languages. These agents answer scam calls, respond to phishing texts, click on malicious links, and infiltrate scam group chats around the clock.
The company, originating from research at Macquarie University in Sydney, was founded in 2024. The core task of the agents is to buy time and extract information such as bank card numbers, cryptocurrency wallets, fake websites, phone numbers, and impersonated brands, which is then packaged and provided to banks, telecoms, and government agencies for blocking. The company claims to have completed over 2.5 million conversations, occupying scammers' time equivalent to about 17 years of work, and has extracted over 250,000 actionable intelligence pieces.
Clients include Commonwealth Bank of Australia and TPG Telecom. TPG has forwarded nearly 590,000 suspicious calls to its agents; Commonwealth Bank reported a 76% decrease in reported scam losses; another unnamed Australian bank used its outbound agents to disrupt 22 scam activities. The agents utilize cutting-edge models and proprietary fine-tuned models, with training data sourced from actual conversations and founder Dali Kaafar's research at Macquarie.
In August 2026, Apate.ai completed an $8.15 million seed round led by Lobby Capital, with participation from OIF Ventures, Investible, Concept Ventures, and Baobab Ventures, bringing total funding to nearly $10 million. The company has re-registered as a Delaware corporation and is advancing global expansion, planning to open an office in London next quarter and strengthen its presence in the U.S., with co-founder and Chief Product Officer Peter Eckermann relocating to the U.S. to oversee North American operations.
This model represents an event-driven proactive anti-fraud approach, reducing institutional losses by consuming scammers' time and extracting real-time intelligence; beneficiaries include banks, telecoms, and government clients (who can preemptively block funding channels), while the pressured party is the scam groups that rely on high call volumes, as their effective working time is significantly occupied by AI agents.
Source: Public Information
ABAB AI Insight
Dali Kaafar, founder of Apate.ai, analyzed scam call content using machine learning and natural language processing during his time at Macquarie University. His experience of actively delaying a scam call for 44 minutes during a family picnic transformed academic research into a commercial product. The company previously received early funding from OIF Ventures and was selected for relevant startup competitions.
In terms of capital, the $8.15 million seed round (led by Lobby Capital) is aimed at deepening the platform and international expansion, while the company has also relocated its registration to Delaware to support the U.S. market and dispatched co-founders to North America. The motivation stems from the expanding scale of global scam losses (with some estimates exceeding $1 trillion annually), as traditional post-fact detection tools have limited effectiveness, and proactive intelligence can be converted into procurement contracts for banks and telecoms.
Drawing parallels to early honeypots and proactive defense systems in cybersecurity, as well as the call diversion mechanisms some telecom operators use for high-risk calls, Apate is currently in the phase of expanding from local pilots in Australia to global institutional clients, productizing "victim simulation" through a large fleet of agents.
Essentially, this represents a technological replacement: traditional anti-fraud relies on user reporting and post-fund freezing, which is being partially replaced by real-time AI agents actively extracting intelligence. The mechanism is that the cost of scams primarily consists of human time, and AI agents can infinitely replicate delaying and questioning behaviors, allowing the defenders to shift from passive response to proactive intelligence collection, thereby converting the attackers' time costs into actionable data for institutions.
ABAB News · Cognitive Laws
- Time is the scammer's ammunition; delaying disarms them.
- Intelligence precedes loss; simulation precedes interception.
- Once scaled, the cost of human resources approaches zero.