Australian startup Apate.ai closes US$8.15 million seed round, using 350,000 AI bots to fight scams

Australian company Apate.ai has just closed a large $8.15 million seed round led by Lobby Capital, with participation from OIF Ventures, Investible, Concept Ventures and Baobab Ventures, bringing its cumulative funding to nearly $10 million. Previously, in August 2025, Apate had also completed a A$2.5 million seed round led by OIF Ventures with participation from Investible.

Image source: Australian company Apate.ai's official website

This business began with a 44-minute phone call during a picnic

In November 2021, Dali Kaafar, a cybersecurity professor at Macquarie University in Sydney, was enjoying the outdoors with his family. His phone rang with an unknown call; the person on the other end claimed to be from a well-known tech giant and, using fluent scripted lines, tried to scam him.

Kaafar, who researches cybersecurity, did not hang up immediately. Instead, he stalled the scammer on the phone, dragging the call out for a full 44 minutes. That picnic interlude did not end there; Kaafar later turned the incident into a business. He founded Apate.ai in 2024, using large-model-driven AI to impersonate ordinary people's technological anxiety and cognitive slowness, keeping scammers online while extracting collection account numbers, wallet addresses and phishing domains.

According to company disclosures, its AI has completed about 2.5 million conversations, consuming a cumulative amount of scammers' time equivalent to 17 years. Separately, foreign media citing Apate founder and CEO Dali Kaafar reported that the Apate platform has about 350,000 AI bots, has been adopted by banks and has won support from telecom companies. Apate has collected more than 250,000 pieces of information about scammers in real time, including scam URLs, accounts used to receive and transfer illicit funds, and bank account details.

Cao Xiao, chief expert at the China-style Modernization Research Institute of Shanghai University of Finance and Economics, believes that the AI anti-fraud contest is ostensibly a contest of technology but in essence a contest over how quickly national governance systems can answer three institutional questions: "who has the right to deploy deceptive AI, by what standards intelligence is accepted, and to whom those wrongly harmed can turn for compensation." Whoever establishes rules and institutions first can turn this round of technological variables into a historic head start in modernizing anti-fraud governance.

Who is paying for Apate.ai?

According to media reports, Apate's clients include Commonwealth Bank of Australia, Australian telecom operator TPG Telecom, and other utility companies, telecom operators and government agencies.

Cao Xiao, chief expert at the China-style Modernization Research Institute of Shanghai University of Finance and Economics, believes that in the past anti-fraud efforts were mostly "counting after the fact": scammers strike first, victims report, and blacklists are updated afterward. Traditional anti-fraud faces three problems: first, it is always half a step behind, and blacklists can never keep up with scammers' changing tactics; second, costs are asymmetric, with defenders spending huge manpower and resources on screening while scammers try things at almost zero cost; third, it is highly passive, relying on victims to report, so when a scam fails and no one reports it, defenders have difficulty knowing about it.

Cao Xiao said the value of Apate's model lies in addressing these three pain points at once: AI impersonates victims, in turn consuming scammers' time and manpower and pushing costs onto the scammers; before money is transferred, it uncovers leads such as collection accounts and phishing URLs; and what used to be only a few volunteers willing to "tease scammers" is now turned by AI into a scalable business.

Apate co-founder and chief business officer Brad Joffe revealed that TPG has transferred nearly 590,000 calls to Apate's AI agents, while the scam losses publicly reported by Commonwealth Bank of Australia fell by 76%.

From "settling accounts after the fact" to "gathering intelligence beforehand," Cao Xiao believes this is a genuine paradigm expansion in anti-fraud theory, namely a shift from "statistics after losses occur" to "intelligence before losses occur." The people chatting with the AI do not pay; those who pay are the institutions concerned with "what leads can be left after the conversation." For the first time, scam intelligence has become a commodity that can be priced and traded. Anti-fraud is transforming from a compliance cost center into an intelligence-driven proactive defense system, and this shift is irreversible.

However, Cao Xiao cautions that one must remain clear-headed about whether intelligence obtained through such "proactive extraction" is reliable. Its difference from traditional transaction-data risk control is not merely a matter of accuracy, but that the data source and generation method are fundamentally different. For example, once scammers know there are AI honeypots, they can deliberately feed fake accounts and fake URLs, and the cost of misleading is very low. Moreover, whether an account is a scammer's collection account is hard to confirm or refute before funds flow in; collection accounts and phishing links may also become invalid within hours and are not easy to detect in time.

Cao Xiao suggests establishing a tiering mechanism for such intelligence, mandating multi-source cross-verification, setting a handling gradient of "slow down only, do not block," and attaching a decay time window and automatic invalidation mechanism to each piece of intelligence. Treating conversational intelligence as grounds for blocking on par with transaction data is the biggest financial risk of this model; treating it as a low-confidence lead that triggers enhanced review is its correct positioning in pre-event risk control.

Where did the $8.15 million in funding go?

In August this year, Apate completed an $8.15 million seed round. Apate co-founder and Chief Business Officer Brad Joffe revealed that Apate has now raised nearly $10 million in total, while the team has only 20 people.

Next, Apate also plans to expand into overseas markets. According to company disclosures, the company will open a new office in London next quarter to serve European customers while strengthening its U.S. business. Co-founder and Chief Product Officer Peter Eckermann will relocate to the United States to lead North American expansion. Apate's technology and research team will expand and remain in Sydney. A parallel move was completed on August 25, when the entity was redomiciled and changed into a Delaware-registered U.S. company to support its global expansion.

There was another similar financing event this year. Also in August this year, Visa announced the acquisition of Israeli anti-fraud company BioCatch for $2.4 billion in cash.

Image source: Visa official website

BioCatch was founded in Israel, and its technology platform detects fraudulent behavior in real time by analyzing thousands of application, behavioral, device and network signals. The signal dimensions collected include keystroke rhythm, touch gestures, touchscreen pressure and device operation patterns, to distinguish genuine users, scammers and bots. The company disclosed that it currently protects 1.8 billion devices and 760 million users worldwide, serving more than 350 bank customers in 21 countries, including more than 100 major global banks.

At the policy recommendation level, Cao Xiao proposed deploying a national pilot for proactive AI anti-fraud perception capabilities, implementing an "authorization list + full-process traceability" approach, and suggested that the Ministry of Public Security take the lead, together with the Ministry of Industry and Information Technology, the People's Bank of China and the Cyberspace Administration of China, to select 2 to 3 cities in provinces with a high incidence of telecom and online fraud and key border directions to carry out pilots, authorizing qualified enterprises to deploy AI decoy systems under police supervision.

"The pilot can implement three rigid constraints: deployment entities, target scope and script boundaries are all included in the authorization list; all interaction records are stored on blockchain, fully traceable and auditable; intelligence outputs directly enter the law enforcement lead pool, and any form of commercial circulation is prohibited. The pilot period is two years, and once mature it can be rolled out nationwide, forming a proactive anti-fraud capability system unique to China in which 'public authority leads and enterprises provide the capability,'" Cao Xiao said.

Cao Xiao judges that in the next five years, AI's reshaping of risk prevention and control will far exceed the single field of anti-fraud. The window period for new scenarios is about three to five years, and risks will not wait for institutions to be ready before appearing; the speed of deployment is itself part of risk control. Ranked by maturity, the following five new implementation scenarios are worth laying out in advance.

First, real-time on-device protection. Google DeepMind has partnered with Singapore's Government Technology Agency to launch an on-device scam detection model based on fine-tuning a lightweight model.

Third, intelligent tracking of fund chains. Graph analysis of abnormal fund flows across banks, payment institutions and virtual currencies.

Fourth, agent identity governance. When AI agents begin to execute transfers, contract signing and customer service operations on behalf of people, "how to prove that an operation instruction comes from the person's true intent" will become a foundational issue in financial risk control and will give rise to entirely new industries such as agent identity authentication and intent confirmation protocols.

Fifth, cross-institutional intelligence federation. Supported by privacy computing technologies such as secure multi-party computation and federated learning, banks, telecom operators and internet platforms can share risk signals without exchanging raw data, breaking the long-standing deadlock between "data silos" and "privacy protection." This is precisely a point of leverage for China's data factor market-oriented reform.