The Perfect Victims: How Sydney Startup Apate Sends 350,000 AI Agents to Get Scammed So You Don’t

When Dali Kaafar got a call from an unknown number during a family picnic, he did what any curious father of young children might do: he decided to entertain the kids. He recognised the scam within a sentence or two, but instead of hanging up, he played along. He kept the fraudster talking for 44 minutes while the children laughed. Then he did a different kind of arithmetic. Those were 44 minutes in which the scammer had not reached a real victim.

That picnic moment became a company. Apate.ai, spun out of Macquarie University in Sydney in 2024, now runs one of the stranger defence operations in cybersecurity: a fleet of AI agents whose entire job is to get scammed, convincingly, at scale. And investors are paying attention. In August the startup closed an $8.15 million seed round led by Lobby Capital, with participation from OIF Ventures, Investible, Concept Ventures and Baobab Ventures, bringing its total raised to just under $10 million. According to TechCrunch, which covered the round, the company plans to put the money into product work and international expansion, including a London office and a bigger presence in the United States.

Turning scam-baiting into an industrial operation

Scam-baiting is not new. For years, YouTubers and hobbyists have filmed themselves wasting scammers' time with fake personas. Apate's twist is to automate the whole thing with conversational AI and aim it at gathering intelligence rather than making content. Thousands of AI agents, built on frontier models combined with systems trained internally on past scam conversations, answer suspected scam calls, reply to dodgy texts and emails, follow phishing links and even join suspicious social media groups.

Each agent plays a character: different ages, genders, accents and languages. One cannot find its reading glasses. Another has to check with the son-in-law first. A third is very keen to invest but keeps mishearing the account number. The point is to hold the fraudster's attention long enough to extract the things scammers normally reveal only after they have a victim: bank account numbers, crypto wallet addresses, phone numbers, malicious websites, the brands they are impersonating and the scripts they are running.

According to the company, the fleet now numbers around 350,000 agents and has logged some 2.5 million scam conversations, which Apate describes as roughly 17 years of fraudsters' time burned on fake victims. Customers see the extracted intelligence through a dashboard and can also hand agents specific investigative assignments, for example tracking down an attacker's account number or wallet.

Real customers, real diversion numbers

The model only works if someone answers the phone on behalf of real targets. In Australia, telecom operators route suspected scam calls directly to Apate's agents, where an AI "victim" picks up. TPG Telecom has diverted nearly 590,000 calls to the service, according to reports. Commonwealth Bank of Australia is among the banks buying the intelligence the bots extract: the mule accounts, the crypto wallets, the hours of operation that a scammer reveals mid-pitch in the belief that he has a live one on the line.

That pipeline from deception to data is what separates Apate from the hobbyist tradition. A YouTuber wasting a scammer's afternoon is entertainment. A telco diverting half a million calls a year into a structured extraction operation is a defence layer, feeding real-time information back into fraud teams before other customers get hit.

Why the timing matters

Apate's pitch has landed at a moment when scammers themselves are going AI-first. The US Federal Bureau of Investigation received more than 22,000 complaints of AI-related crimes in 2025, with losses of about $893 million, TechCrunch reported this month. More than $632 million of that came from investment scams, in which fraudsters use AI to run conversations and deepfake the voices and faces of celebrities, executives and trusted figures. It is a strange arms race: the same generation of voice and language models that lets criminals clone a grandchild's voice also lets a Sydney startup mass-produce the perfect victim.

Kaafar, a professor at Macquarie University where the research behind Apate began, put it plainly to WIRED: "What we really like to think is that we're building the perfect victims for scammers. A minute that a scammer is talking to a bot or an agent is a minute where you're probably saving hundreds, if not thousands of possible people being reached out to by that exact same scammer."

From picnic prank to Startup Battlefield finalist

The company was named after Apate, the Greek goddess of deception, which is either charming or foreboding depending on your view of AI agents. It was selected as a finalist in TechCrunch's Startup Battlefield 200, the annual startup competition held at Disrupt in San Francisco, giving it visibility with exactly the kind of customers and partners it now wants to court abroad. With 20 staff and the fresh seed money, the plan is to go international rather than grow domestically, starting with London and a larger US footprint.

The bigger question is what happens when scam-baiting scales up everywhere. If every major bank and telco starts routing suspected scam calls into agent farms, fraudsters will eventually build counter-detection, which pushes the cat-and-mouse game one level deeper. But for now, Apate is betting that the asymmetry favours the defender: it is cheap to spin up another gullible grandmother persona, and expensive for a scammer to figure out which "victims" are worth the effort.

Frequently Asked Questions

How do Apate's AI agents actually work?

The company runs a fleet of AI agents that pose as scam victims across phone calls, text messages, emails and online platforms. Telecom operators divert suspected scam calls to the agents, which adopt distinct personas with different ages, accents and temperaments, keep fraudsters engaged and extract intelligence such as bank accounts, crypto wallets and malicious websites that is passed to banks, telcos and agencies.

Which companies are using Apate's technology?

According to reports, Apate's customers include banks, telecom companies, government agencies and utilities. Publicly named customers include TPG Telecom, which has diverted nearly 590,000 suspected scam calls to the service, and Commonwealth Bank of Australia, which uses the extracted intelligence for its anti-fraud work.

How much money has Apate raised?

Apate raised $8.15 million in a seed round in August, led by Lobby Capital with participation from OIF Ventures, Investible, Concept Ventures and Baobab Ventures. According to TechCrunch, the company has raised just under $10 million in total and plans to use the funding for product development and international expansion, including a London office and a bigger US presence.