AI disabled
Human-only workflow for sensitive or policy-restricted projects.
AI GOVERNANCE
Client-controlled AI, not silent AI.
Projects can disable AI entirely, restrict processing to private enterprise models, use approved providers, or use a client-provided API key.
Private AI
Designed for environments where model access and retention must be tightly controlled.
Client API key
Enterprise customers can route AI processing through their own approved provider configuration.
Editorial review: Tasheel Express Language Operations Team · Last reviewed: 15 August 2026.
Practical guidance
Controls that help choose the right workflow
How AI is selected for a project
AI is not treated as the default for every file. The project is first classified by content type, confidentiality, intended use, regulatory sensitivity, language pair and review requirements. Human-only, private AI, approved enterprise AI or a client-controlled provider can then be selected deliberately.
Human accountability remains explicit
AI output is not presented as a substitute for professional responsibility on high-risk content. Legal, medical, official, safety-critical and brand-sensitive material can require professional review, terminology checks and final human approval before delivery.
Data minimization and retention
The intake workflow records the project’s AI preference, data region and retention preference. Sensitive fields can be redacted before automated processing where appropriate. Production data-residency guarantees depend on the infrastructure actually configured for the client.
What we do not claim
Tasheel Express does not claim that every AI output is certified, legally valid or suitable for filing without review. Certification and receiving-authority requirements are confirmed separately from the technology used to assist translation.