Pomegra
Apate.AI raised A$11.4M from Lobby Capital to scale adversarial AI bots that impersonate scam victims and extract live fraud intelligence for banks and telcos.
- Lobby Capital led Apate.AI's A$11.4M seed on August 31, 2026, with OIF Ventures, Investible, Concept Ventures, and Baobab Ventures co-investing.
- Apate's AI bots engage live scammers in real-time voice and text conversations, harvesting fraud network intelligence for client banks and telcos.
- The company is entering North America and Europe, with co-founder Peter Eckermann relocating to the US and a London office opening to serve Europe.
Sydney Startup Turns Scammers' Own Playbook Against Them
Apate.AI, a 2024 Macquarie University spin-out, closed an A$11.4 million seed round on August 31, 2026. Lobby Capital, a Silicon Valley firm, led the raise. OIF Ventures, Investible, Concept Ventures, and Baobab Ventures participated. The round follows a A$2.5 million seed closed roughly twelve months earlier, also anchored by OIF - a funding pace that suggests enterprise traction arrived ahead of schedule for a company still under two years old.
What Does Apate.AI Actually Build?
The product is a fleet of conversational AI agents designed to pass as prospective scam victims. When a fraudster initiates contact - through a phone call, SMS, or messaging app - Apate's agents respond in real time, sustaining the exchange long enough to document phone numbers, scripts, account references, and behavioral fingerprints. That intelligence routes back to client institutions: banks, telecommunications carriers, and government agencies. The system simultaneously wastes the scammer's time and generates proprietary fraud-network data that passive detection systems cannot produce.
Co-founders Dali Kaafar, Peter Eckermann, and Brad Joffe originated the concept at Macquarie's Applied AI Research Centre. The company received an Australian Economic Accelerator seed grant from the federal Department of Education, which supported the move from academic prototype to commercial product before private capital entered. Adversarial techniques - honeypots, phone traps - have existed in security for decades. Apate's differentiation is scale and integration: AI agents running many simultaneous live interactions across voice and text, with outputs structured for direct ingestion into institutional fraud-management platforms.
Why Is This Round Still Labeled Seed After A$2.5M Already Closed?
Calling the A$11.4 million round seed rather than Series A lets Apate defer a formal post-money valuation until the company can demonstrate repeatable enterprise revenue across institution types. Valuation remains undisclosed. Both rounds carry seed designations because the company has not yet priced itself publicly - a deliberate negotiating posture ahead of a Series A, though it also means investors in both rounds hold instruments against an unpriced asset.
The Lobby Capital lead reflects a shift in commercial focus toward North American institutional buyers. Eckermann will relocate to the US to lead that expansion. A London office is planned to anchor Europe. Engineering and research remain in Sydney, a structure common to Australian deep-tech exporters but one that introduces coordination overhead as headcount spreads across three time zones simultaneously.
Regulatory Pressure Is Pushing Banks Upstream
Australian scam losses reached A$2.74 billion in 2023 and remained elevated through 2025. Regulatory frameworks in Australia, the UK, and the US now hold financial institutions and telcos partly accountable for customer scam losses, forcing demand for tools that disrupt fraud before a transaction clears rather than flagging suspicious activity after funds have moved. Apate's product sits at exactly that upstream point.
The broader fraud-intelligence sector has attracted significant capital, mostly directed at detection and authentication. Few companies are deploying live adversarial engagement at institutional scale. If Apate's agents generate scam-network data structurally unavailable to passive vendors, the company has a credible path to a proprietary intelligence layer, one that could support data-licensing revenue alongside software subscriptions and differentiate it from commoditizing fraud-detection providers.
How Do Legal Constraints Shape Deployment in New Markets?
Operating undisclosed AI agents in live conversations with individuals - even fraudsters - creates jurisdictional exposure. Australia's telecommunications interception laws, US state and federal wiretapping statutes, and UK communications legislation each impose different thresholds for what constitutes lawful interception or recording. The company's government grant history and academic origins suggest regulatory awareness, but the mechanics of legal coverage for live voice deployments across multiple markets have not been disclosed publicly. Each enterprise deal likely requires its own legal structure, which adds friction to sales cycles, particularly in the US where state-by-state variance is significant.
Data handling carries parallel obligations. Bot conversations may incidentally capture personal data, creating obligations under GDPR in Europe, US state privacy laws, and Australia's Privacy Act - obligations that will require careful scoping in each institutional contract.
Outlook
Apate.AI enters a fragmented fraud sector with a product category mostly its own. The A$11.4 million gives the company an estimated twelve to eighteen months of runway to land anchor enterprise contracts across three geographies before a Series A becomes necessary. The critical variable is procurement pace: financial institutions move slowly, and Apate is targeting banks, telcos, and government agencies simultaneously across the US, UK, and Australia. Revenue visibility from the earlier A$2.5 million round - which the company has not detailed publicly - will set the valuation terms for what comes next.
