AI bots posing as scam victims have wasted 17 years of fraudsters’ time. Now this startup is expanding to the US

“Traditional approaches to scam defence rely on after-the-fact intelligence,” Apate.ai says on its website. “Our approach is different. Rather than reacting after the damage is done, we put our armies of lifelike AI bots right in the firing line of scammers, disrupting their operations in real-time and stopping scam at the source.”

The technology focuses on collecting data and buying time, two issues that have vexed fraud fighters.

“Our conversational AI is trained on extensive scam data, enabling it to respond naturally and keep scammers engaged for prolonged periods,” the company added. “This ensures that fraudsters waste valuable resources while we gather the intelligence needed to shut down their operations.”

Apate.ai says its AI agents have wasted 17 years of scammers’ time across 2.5 million conversations, according to TechCrunch. Its agents have the capacity to answer scam calls, messages, emails and can follow phishing links and join social media groups. The company’s approach could be welcome news for the fraud-prevention sector.

Data from the FBI’s Internet Crime Complaint Center shows that in 2025, more than 1 million complaints were filed and nearly $20.9 billion was lost to cybercrime — up 26% from 2024. Additionally, the Federal Trade Commission reported $15.9 billion lost to consumer fraud in 2025.

Gathering the fraud information that matters

Apate.ai, which is rolling into the U.S. market after an $8.15 million funding round in August, lays out an intriguing case for its “armies of adversarial conversational AI agents,” as it notes on its website.

“Most of the industry works after the fact,” Brad Joffe, co-founder and chief commercial officer, told TechCrunch. “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.”

Cybersecurity experts say Apate.ai is onto something, giving fraudsters a foe they’ve never encountered before.

“There’s real potential here because we’re shifting from reacting to fraud after it happens to identifying threats before consumers become victims,” Mary Ann Miller, fraud executive advisor at Prove, a digital identity services company, told Moneywise.

Miller noted that AI agents that engage scammers could help uncover fraudulent accounts, tactics and patterns that financial institutions can use to strengthen their defenses. “But the real measure of success isn’t how much time scammers waste talking to bots,” she said. “It’s whether that intelligence translates into fewer successful scams and less money lost by consumers.”

A relatively new idea in fraud fighting, AI-powered scam baiting could finally give banks and consumers an edge over increasingly sophisticated scam artists.

“One of the biggest challenges in fraud prevention is that criminals are getting more sophisticated and constantly adapting and financial institutions are often left responding to new tactics after an incident has occurred,” Miller said.

Using AI to proactively engage scammers and identify emerging threats could give financial institutions valuable intelligence earlier in the process. “The opportunity is connecting those insights to real-time fraud detection, identity verification and transaction monitoring so suspicious activity can be stopped before money moves,” Miller added.

Clients report a big reduction in fraud, but there are caveats

Apate says its customers are reporting impressive fraud-reduction numbers, according to TechCrunch. For example, the Commonwealth Bank of Australia has seen a 76% decline in publicly reported scam losses, while Joffe said another Aussie bank has used its agent to disrupt 22 scam campaigns.

Could the numbers be too good to be true? Cybersecurity experts say that’s possible.

“Fraud loss estimates are notoriously difficult to attribute accurately — reporting rates change, fraud patterns change seasonally and economic conditions impact both fraud volume and victim reporting,” Saugat Nayak, a data science and risk analytics specialist and a former analyst at Deloitte Consulting, told Moneywise.

Nayak said a 76% fraud reduction is a stark number and requires looking at the baseline period, the measurement methods and what else happened at the same time. “That said, if the technology really does impair fraudulent infrastructure quality and speed, directional progress is feasible,” he said.

Other fraud-prevention professionals say the real question is whether American consumers might expect a similar reduction.