How Apate’s undercover agents work
TechCrunch reports that Australian startup Apate.ai deploys AI agents that pose as potential victims in conversations with suspected scammers. They can engage through calls, messages, email and other online channels. The goal is to keep a suspected fraudster talking long enough to learn about the operation before a real person is harmed.
The idea grew from founder Dali Kaafar’s experience keeping a scam caller on the line for 44 minutes, according to TechCrunch. Apate now offers that kind of engagement at much larger scale to organizations trying to spot and interrupt scams earlier.
What banks and telecoms can learn
Apate says its agents try to surface bank-account and crypto-wallet details, phone numbers, malicious websites and the brands criminals impersonate. It supplies that intelligence to customers that include Commonwealth Bank of Australia and TPG Telecom. Those organizations can use the leads in their own fraud investigations and defenses.
This is different from simply blocking a phone number after a complaint. The company’s pitch is to identify a campaign and its infrastructure while the conversation is still underway. Whether a given lead results in a stopped scam depends on what the customer can verify and act on.
The reported scale—and the limits of the evidence
Apate told TechCrunch that its agents have had 2.5 million conversations and occupied the equivalent of 17 years of scammers’ time. These are company-provided figures, not independently verified measures of prevented fraud. The number of conversations alone does not show how many criminals were identified or victims protected.
TechCrunch also reported customer examples, including a reduction in publicly reported scam losses at one bank. That change cannot be attributed to Apate alone without a controlled comparison and fuller information about other anti-fraud measures. AI decoys may be useful, but they are one part of a broader defense.
Why this matters—and where to watch and listen
Scammers increasingly use AI to scale convincing outreach. Apate’s approach applies AI on the defensive side: engage suspicious contacts, gather leads and give institutions a chance to respond sooner. It also raises practical questions about accuracy, oversight and how reported success should be measured.
Watch the Daily AI Chat Short or listen to the full podcast episode. Source: Ram Iyer’s TechCrunch report, October 14, 2026.
