AI Governance & Data Handling
iGears builds AI systems for clients and uses AI in its own work. This page states the rules we hold ourselves to: client data stays inside agreed boundaries, humans review what ships, and we do not oversell what AI can do.
Client data boundaries
Client documents, databases and business information are used only for the engagement they were provided for. When an AI solution processes client data, the processing boundary — cloud, region, or fully on-premises — is agreed in writing before anything is built. Private and on-prem deployment is a standard offering, not an exception.
No training on client data
We do not use client data to train models for other clients, and we configure third-party AI services so that client content is excluded from provider training wherever the provider offers that control.
Human review
AI-assisted work — code, content, analysis — is reviewed by the responsible engineer or consultant before it reaches a client. Automated workflows we deliver include human checkpoints wherever an action is irreversible or customer-facing, unless the client explicitly specifies otherwise.
Honest claims
We tell clients what AI can and cannot do for their case before they spend money. Fixed-scope assessments exist precisely so that ambition is tested against evidence early. We do not present AI-generated material as human work where the difference matters.
Security basics
Access to client systems follows least-privilege, credentials are managed per-project, and AI tooling runs under the same security review as any other software we operate.
Questions
If your organisation needs a specific governance arrangement — data residency, audit trails, model choice, procurement documentation — ask. This is normal work for us. Get in touch.
Last updated: 9 August 2026. Questions about this page? Email info@igears.ca.