Apate.ai uses AI agents as honeypot “victims”: the bots answer scam calls, reply to suspicious messages, follow phishing links and enter fraud groups so criminals keep talking while investigators collect actionable intelligence. The approach is proactive threat intelligence, not merely a number-blocking service.
What Apate.ai is and why it matters
Apate.ai is an Australian fraud-intelligence company founded in 2024 from research at Macquarie University in Sydney. Its public products are Apate Voice, Apate Text and Apate Insights. Together, they are intended to reveal how a scam operates before a genuine customer sends money or shares credentials.
As an Amazon Associate I earn from qualifying purchases.
The timing matters because fraudsters can replace a blocked phone number, domain or social account quickly. Conventional defenses often see the number, message or transaction only after it has reached a potential victim. Apate instead creates controlled identities that scammers can discover and targets the scammer’s infrastructure directly.
How Apate “scams the scammers”
Apate uses a honeypot model. Dedicated phone numbers and controlled digital identities are made discoverable to criminals. When a scammer calls or sends a message, an AI agent—not a real customer—conducts the conversation.
#1 Best Overall
- ALL-IN-ONE SCAM DETECTION – Texts, emails, videos, and QR codes all get checked automatically. Sorting real from fake stops being your job.
- KEEP SCAMMERS OUT OF YOUR WALLET – Every click is no longer a gamble. Our scam detection spots suspicious texts, email scams, SMS phishing, and fake alerts before you click.
- QR CODE SCANNING – Point the app at any code and see where it actually leads before you scan it.
- DEEPFAKE DETECTION – When a video sounds like someone you know but isn't, you hear it from us first.
- ON-DEMAND CHECKS – Got a message you're unsure about? Run it through the app and know in seconds, wherever it came from.
The agents are designed to appear credible by varying apparent age, gender, accent, language and cultural cues. An investigation can assign a mission, such as obtaining a bank-account number or a cryptocurrency-wallet address. The bot can ask questions, delay decisions and follow the scammer’s instructions long enough to expose the wider operation.
- Attract contact: a dedicated number or digital identity enters the pool of targets that scammers can find.
- Hold the conversation: an agent answers a voice call or continues a text exchange while behaving like a plausible victim.
- Follow the trail: the agent can open phishing links, inspect messages and participate in relevant online groups under controlled identities.
- Extract indicators: systems record phone numbers, malicious websites, impersonated brands, scripts, payment routes, mule infrastructure and links between campaigns.
- Deliver intelligence: the resulting records feed alerts, analytics, investigations, blocking decisions and takedown or compliance processes.
What the three public products do
Apate Voice
Apate Voice handles conversational phone calls. Apate says its agents can keep a scammer engaged for up to 50 minutes and can run many conversations at once. A long call can expose the script, the requested payment path, a supposed support number or the next person in a criminal workflow.
Apate Text
Apate Text works across WhatsApp, Telegram and SMS. It can also search forums, investment groups and social platforms for scam activity, allowing investigators to connect a direct message with the communities and accounts promoting it.
Apate Insights
Apate Insights turns voice and text interactions into dashboards, alerts and analytics. The company positions this layer as the operational interface for analysts who need to prioritize threats, connect related indicators and export records for investigations or reporting.
What intelligence the agents can uncover
- Mule networks: payment routes, transaction patterns, recruitment tactics and infrastructure used to move stolen funds.
- Impersonation campaigns: relationships among phone numbers, scripts, fake personas, accounts, channels and the brands being imitated.
- Phishing operations: malicious links, landing pages and campaign connections that can support earlier blocking.
- Evidence for action: timestamped, structured and exportable records that provide an audit trail for compliance teams, takedowns and reports.
The practical benefit is correlation. A single suspicious number may look disposable; a graph linking that number to a script, a domain, a payment account and several social profiles can identify a campaign that should be disrupted.
Apate versus blocking and manual scam-baiting
| Approach | Channel coverage | Conversation capacity | Intelligence produced | Operational timing |
|---|---|---|---|---|
| Conventional blocking | Usually focused on known phone numbers, domains or accounts | Ends when the indicator is blocked or replaced | Primarily an indicator for a blocklist | Reactive; often follows a report or attempted contact |
| Manual scam-baiting | Any channel a human investigator can reach | Limited by staff time and safety | Can be detailed but varies by investigator and case notes | Small-scale and slower to repeat |
| Apate agents | Voice, WhatsApp, Telegram, SMS, forums, investment groups and social platforms | Apate says voice agents can sustain calls for up to 50 minutes | Structured links among numbers, sites, scripts, personas, payment routes, mule infrastructure and campaigns | Designed for proactive engagement, near-real-time alerts and operation at large scale |
Apate does not eliminate blocking or transaction monitoring. It supplies earlier intelligence for those controls. Telcos can use discovered indicators to block contacts, banks can investigate payment paths, and government or enterprise teams can coordinate reports and takedowns.
Rank #3
Who is using Apate
TechCrunch reported customers including Commonwealth Bank of Australia, TPG Telecom, other utilities and telecommunications companies, and government agencies. Commonwealth Bank separately said Apate had deployed thousands of bots and that hundreds of thousands of scam calls had been diverted since a pilot that began in late 2024. Its announcement described near-real-time alerts and blocks generated from the intelligence loop.
The customer mix illustrates the platform’s intended role: a telco can see the calling infrastructure, a bank can follow payment intelligence, and public agencies can use shared campaign evidence for disruption and reporting. Exact integrations and commercial terms depend on the organization; Apate has not published a universal deployment specification.
Reported results and how to interpret them
The following figures were reported by TechCrunch in 2026 and attributed to Apate, its customers or company executives. They are not independently audited measurements:
Rank #4
| Reported result | Attribution and qualification |
|---|---|
| 17 years of scammers’ time wasted across 2.5 million conversations | Apate figure reported by TechCrunch in 2026 |
| Nearly 590,000 scam calls diverted to Apate agents | TPG Telecom figure attributed to Apate and reported by TechCrunch in 2026 |
| 76% drop in publicly reported scam losses | Commonwealth Bank figure attributed to Apate and reported by TechCrunch in 2026 |
| 22 scam campaigns disrupted | Figure from an unnamed Australian bank, attributed to Apate and reported by TechCrunch in 2026 |
These numbers should be read as reported customer or company outcomes, not as a controlled comparison with organizations that did not use Apate. The broader problem is substantial: TechCrunch reported that the FBI received more than 22,000 complaints involving AI-related crimes in 2025, with about $893 million in losses; more than $632 million was associated with investment scams. Those FBI figures describe the wider AI-crime environment, not Apate’s performance.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why sustained engagement is the core advantage
Every minute an automated agent occupies a fraudster can delay contact with a real target. More importantly, a longer exchange gives analysts more opportunities to identify the criminal’s workflow. A short interaction may reveal only a phone number; an extended conversation can reveal a payment destination, a second-stage operator, a fake investment platform and the language used to recruit money mules.
Recommended Free Tools
That is the distinction behind Apate co-founder Brad Joffe’s description of the system: number blocking and transaction monitoring often happen after a scammer has already reached someone, while Apate aims to obtain information before that harm occurs. CommBank fraud executive James Roberts similarly described the approach as “flipping the script” against scammers who increasingly use AI themselves.
Best Value
Limits, safeguards and questions for adopters
Apate’s agents use fabricated identities and controlled contact points, but any organization deploying such a system still needs clear rules for data handling and escalation. Before deployment, a fraud or security team should establish:
- which phone numbers, accounts and online identities are controlled by the organization;
- what information agents may request or receive during a mission;
- how recordings, messages, links and account details are timestamped and retained;
- which indicators can trigger an automatic block and which require analyst review;
- how evidence is shared with banks, telcos, platforms, regulators or law-enforcement partners;
- how investigators distinguish a related campaign from coincidental reuse of a number, script or brand.
Apate is therefore best understood as a threat-intelligence and fraud-operations platform. It can strengthen blocking, monitoring and takedown workflows, but it is not a consumer app that guarantees a person will never receive a scam message.
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.
