The first platform built to classify insider intent before it becomes an incident.
irari AI analyses psychophysiological data to classify intent, giving organisations an objective view of insider risk before any harm is done.
The problem
Security spending rarely reaches the threat that costs the most.
Most organisations invest heavily to defend against outside attackers. But the costliest breaches usually come from within, and almost nothing in the security stack is built to see them coming.
Source: Verizon Data Breach Investigations Report, 2022. The information-compromised split is illustrative, as although insiders account for only ~20% of breaches, the report finds they compromise over 5× more information per breach than outsiders.
Breaches caused by malicious insiders are the costliest of any attack type organisations face, and the global average cost of a single insider threat incident is
Source: IBM Cost of a Data Breach Report, 2026.
Most insider threat incidents never get reported or prosecuted at all, meaning the true scale and impact of the problem is much higher.
Source: Claycomb et al., 2012.
How it works
Built for strategic foresight, not hindsight.
irari AI is a signal acquisition and AI classification system that assesses insider threat intention before an incident occurs. Through a short digital assessment, it captures multimodal temporal signals, including response latency, gaze, and cursor dynamics, and converts them into an objective, auditable risk score for authorised decision-makers.
The platform integrates cognitive measurement science, signal processing, and proprietary AI, grounded in scientific (doctoral) research into the psychology of intention and validated across independent experiments. It supports, not replaces, human judgment, works alongside existing security tools, and enables strategic insider threat management.
To help organisations safeguard their people, information, assets, and reputation, by making insider threat assessment holistic, fair, and proactive.irari's mission
Team
Deep expertise, proven collaboration.
- EducationPhD, Psychology — University of Otago; MSc (cum laude) — University of Twente
- FocusProduct management, experimental research, data collection & analysis
- RecognitionKiwiNet Emerging Innovator, 2024–25
- Track recordSecond-time founder; Ex-Product Owner at JustEatTakeaway
Three years of doctoral research into the psychology of insider threat, building on earlier research in the Netherlands that identified a critical gap in existing approaches. The resulting IP is wholly owned by the company. Recognised as a KiwiNet Emerging Innovator for its commercial potential and bringing second-time founder experience from co-founding an AI hardware company.
- EducationBSc, Computer Science
- FocusTransformer architectures, deep learning systems
- StatusCompany's first engineer, upon funding
- Working relationshipAlready collaborates with the founder at another company
Experienced in building and deploying deep learning systems, including transformer-based architectures and temporal sequence models directly relevant to irari AI. The company’s first engineer, bringing both the technical depth required to build the core platform and an established relationship with the founder, reducing early-stage execution risk.
Get in touch
Let's work together.
irari AI is currently in stealth. For partnership, pilot, or investment conversations, reach out directly. We'd welcome it.