Account takeover can start before the first login
A defensive review of online enrollment, account discovery, recovery, and the controls between a customer record and digital access.
Independent thinking. Practical research.
Tervaq is an AI research and product company. We build intelligent software and work with organizations to understand emerging risks, test critical systems, and turn research into practical results.
Drag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
01 / A global perspective
Fraud rarely fits neatly inside one team’s remit. Identity, account access, customer support, and transaction controls all shape the outcome. Our research follows the workflow across those boundaries.
We examine account abuse, malicious automation, and AI-enabled threats in the context of the people and systems affected. A global perspective means understanding differences in markets and infrastructure, not assuming one control works everywhere.
Our research agendaDrag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
02 / Applied AI
We research how AI reasons, coordinates, and retains useful context. The aim is software that helps people do meaningful work without hiding how it operates.
Drag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
03 / Security Intelligence
Fraud intelligence, AI abuse research, and authorized security assessments. We help teams separate claims from evidence and test the controls that matter.
The work includes enrollment and recovery reviews, investigation of suspicious workflows, and testing of proposed fixes. Reports separate what a source claimed, what was observed, and what an authorized assessment established.
Explore Security IntelligencePaid engagements. Defined scope. Written authorization.
Drag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
04 / Enterprise systems
Context, memory, permissions, and reliable handoffs. We study the architecture that turns an isolated capability into something an organization can operate.
Drag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
05 / From research to products
We build products as well as study systems. Each has a defined purpose, its own development path, and a place in the wider Tervaq portfolio.
Explore the portfolioDrag to rotate. Double-click to reset. Use arrow keys to turn, plus or minus to zoom, and Home to reset. On touch screens, swipe sideways to turn and scroll vertically to continue.
Security research & fraud intelligence
A report should give the receiving team a decision it can make. We scope research around the affected workflow, the evidence available, and the people responsible for acting on it.
Our focus covers the path from establishing identity to gaining access and making sensitive changes. Enrollment, recovery, device trust, and customer-service exceptions need to be understood together.
For suspected fraud, we distinguish the source’s account from behavior that has been directly observed. Where technical validation is needed, the scope, test identities, access, and stop conditions are agreed in writing.
Explore the practiceInvestigations, technical assessments, and remediation validation are paid engagements. Scope, deliverables, and commercial terms are agreed before work begins.
Discuss your research questionFrom the research desk
Notes for the people responsible for protecting accounts, evaluating AI, and deciding whether a control actually works.
Read the journalA defensive review of online enrollment, account discovery, recovery, and the controls between a customer record and digital access.
How to separate source claims from observations, preserve context, and give a security or fraud team a decision-ready evidence brief.
A practical way to assess claims of AI-assisted fraud without confusing model capability, automation, and a demonstrated control failure.
Need a place to start? Use our blank review worksheets and evidence-brief templates in your own approved environment.
Browse resourcesWork with Tervaq
Bring an account-security concern, an unexplained fraud pattern, or an AI workflow that needs closer examination. We’ll discuss the evidence available, the question to investigate, and the scope and fees before commissioned work begins.
Start a conversationResearch partnerships · Assessments · Intelligence retainers