iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Apate.ai turns the scammer-targeting model around: its AI agents pose as potential victims, answer calls, reply to messages, follow phishing links and enter suspicious online groups. While keeping criminals occupied, the system records the phone numbers, websites, scripts, payment routes and relationships needed to disrupt a campaign before it reaches more real victims.
The Australian company is building a proactive fraud-intelligence layer for banks, telecommunications companies, utilities and government agencies—not merely another service that blocks known numbers after a scam has been reported.
What Apate is designed to do
Founded in 2024 from research at Macquarie University in Sydney, Apate.ai operates a honeypot network of controlled phone numbers and digital identities. Those identities are made discoverable to fraudsters. When a criminal calls or sends a message, an AI agent—not a human victim—conducts the conversation.
Recommended Free Tools
The objective is twofold: consume the scammer’s time and obtain operational intelligence. Agents can be given missions such as finding a bank-account number or cryptocurrency-wallet address. The resulting evidence can connect a telephone number to a malicious site, an impersonated brand, a payment route, a script, a mule network or other campaigns.
#1 Best Overall
Apate co-founder and chief commercial officer Brad Joffe describes the rationale this way: “Most of the industry works after the fact. 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.”
How the undercover agents work
- Create a discoverable honeypot. Dedicated telephone numbers and controlled online identities are connected to partner networks and presented in places where scammers may find targets.
- Engage the incoming scammer. A bot answers a call or responds in a text channel, maintaining a credible conversation rather than ending the interaction after identifying a suspicious number.
- Use believable personas. Apate says its agents can vary age, gender, accent, language and cultural cues. That diversity helps the system handle campaigns aimed at different communities.
- Run an assigned mission. An agent may be tasked with eliciting a wallet address, bank details, a link, a recruitment process or other information useful to investigators.
- Extract and correlate indicators. Conversations can reveal numbers, accounts, domains, scripts, personas, payment infrastructure and links between campaigns. Those indicators can then be matched across interactions and channels.
- Deliver intelligence to defenders. Apate Insights provides dashboards, alerts and analytics intended for fraud, security, compliance and takedown workflows.
Apate says a voice agent can keep a scammer engaged for up to 50 minutes. The duration is important because the system is not simply classifying a call; it is trying to delay the criminal while collecting details that can support intervention.
Apate’s three public products
| Product | Primary channel or function | What Apate reports |
|---|---|---|
| Apate Voice | Telephone conversations | Conversational agents operate at scale and can keep scammers engaged for up to 50 minutes, according to Apate. |
| Apate Text | WhatsApp, Telegram, SMS and online communities | Agents respond to text scams and can search forums, investment groups and social platforms for activity. |
| Apate Insights | Intelligence management | Dashboards, alerts and analytics turn interactions into information that teams can investigate, share and act on. |
What intelligence defenders can use
Mule-network investigation
Agents can identify how criminals recruit money mules, where payments are routed and which transaction patterns connect apparently separate operations. Earlier visibility can give a bank or investigative team a lead before a suspicious transfer is completed.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Impersonation and phishing mapping
Linking phone numbers, scripts, personas, accounts, channels and financial infrastructure helps analysts map a campaign instead of treating every message as an isolated event. Malicious websites and impersonated brands become part of the same evidence set.
Rank #3
Compliance and takedown support
Structured, timestamped and exportable records provide an audit trail for reporting, coordinated blocking and requests to remove malicious infrastructure. The practical value is the chain of evidence: who contacted the honeypot, what was requested, which link or account was supplied and how that indicator relates to other activity.
How Apate differs from ordinary scam defenses
| Capability | Conventional number blocking or post-transaction review | Apate’s documented approach |
|---|---|---|
| Channel coverage | Often focused on known telephone numbers, transactions or reported messages | Voice and text interactions, including WhatsApp, Telegram, SMS, forums and social or investment groups |
| Conversation handling | Usually stops or flags an interaction | AI agents sustain a conversation and can pursue a defined mission |
| Intelligence depth | Frequently limited to a number, URL or transaction alert | Can collect scripts, personas, payment routes, mule infrastructure and campaign links |
| Deployment scale | Human analysts or rules process cases one at a time | Thousands of controlled identities and agents can operate concurrently, according to Apate and its customers |
| Time to alert | May follow a complaint, transaction or confirmed abuse | Designed to produce intelligence while the scam is still being run |
| Fraud-workflow integration | Requires separate investigation and reporting steps | Apate Insights supplies dashboards, alerts and exportable records for investigation, compliance and takedown work |
| Cross-sector sharing | Signals can remain inside the organization that received them | Indicators can help banks, telcos, utilities and agencies connect campaigns across sectors |
| Auditability | May depend on call logs or customer reports | Timestamped interaction records are intended to support an evidence trail |
This does not make blocking obsolete. Blocking can stop a known number quickly; Apate’s distinction is engaging the operator first to discover the next number, account, domain or payment route.
Who is using Apate?
TechCrunch reported that Apate customers include Commonwealth Bank of Australia, TPG Telecom, other utilities and telecommunications companies, and government agencies. Commonwealth Bank separately announced that thousands of bots had been deployed and that hundreds of thousands of scam calls had been diverted since its pilot in late 2024. The bank described near-real-time alerts and blocks generated from the intelligence loop.
Free tools Windows power users keep installed
One-click scans. No signup required.
Reported performance figures should be read as attributed claims, not independent audits:
Best Value
- Apate said 2.5 million conversations had wasted a combined 17 years of scammers’ time, as reported by TechCrunch in 2026.
- TPG Telecom attributed nearly 590,000 diverted scam calls to Apate, according to the same report.
- Commonwealth Bank attributed a 76% drop in publicly reported scam losses to the program.
- An unnamed Australian bank attributed 22 disrupted scam campaigns to Apate.
The figures use different measures and organizations, so they are not a single standardized performance benchmark. They indicate how customers describe the program’s operational effect, while the underlying methodology and independent verification have not been published by the organizations in the cited reports.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why this matters as AI-powered fraud grows
The FBI received more than 22,000 complaints involving AI-related crimes in 2025, with about $893 million in reported losses. More than $632 million of that amount came from investment scams. Those FBI figures, reported by TechCrunch, describe the wider threat environment rather than Apate’s performance.
Professor Dali Kaafar, Apate’s CEO and founder, says the system is based on a honeypot strategy operated with telecommunications partners: a growing network of dedicated numbers is designed to be discovered and targeted by scammers. CommBank’s James Roberts called the approach “flipping the script,” arguing that every minute a scammer spends with a bot is a minute not spent targeting an Australian.
What organizations should clarify before deployment
- Evidence standards: Confirm how conversation records, extracted indicators and links between campaigns are validated before they trigger a block or investigation.
- Workflow ownership: Decide which team receives an alert, who approves a takedown and how rapidly a verified indicator reaches fraud controls.
- Coverage expectations: Establish which countries, languages, channels and scam types are included in the deployment; the public descriptions do not define a universal coverage set.
- Measurement: Separate engagement time, diverted calls, disrupted campaigns and reduced losses. These are different outcomes and should not be combined into one score.
- Governance: Set retention, access and export rules for recordings, messages, account details and other evidence collected during undercover interactions.
Public material does not state a standard price, universal technical requirements or a single deployment package. Those details need to be confirmed directly for each bank, telco, government or enterprise-security use case.
The bottom line
Apate is best understood as a proactive fraud-intelligence platform that uses AI agents as honeypot “victims.” Its value is not only delaying a scammer; it is converting that interaction into connected, time-stamped intelligence that defenders can use before a real customer is contacted or a payment is made. Customer-reported results are promising but remain attributed figures rather than independently audited proof.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
