The language of AI adoption.
Plain-English definitions for the terms behind discovering, governing, and proving the value of AI.
Adoption engine
A system that discovers AI usage, governs it, and proves its value in one continuous loop — rather than a static dashboard.
Authorized app
A third-party application granted OAuth access to your org's data, often with AI features that touch sensitive information.
Deep inspection
Optional decryption of AI traffic to capture prompts, responses, and tool calls for governance and evidence.
Desktop agent
A lightweight endpoint program that classifies outbound AI traffic by vendor and emits usage signals.
Drift
A change in an AI system's behavior over time that can quietly break a control between reviews.
Evidence pack
An audit-ready, timestamped bundle of proof that a control is working, exported on demand.
Policy-as-code
Governance rules defined in code and enforced automatically, mapped to frameworks like NIST AI RMF and AIUC-1.
Shadow AI
AI tools and agents used inside a company without IT or security's knowledge or approval.
Tool-call safety
Controls ensuring AI agents only invoke permitted tools, with valid parameters and within limits.
Vendor inventory
A continuously updated list of every AI vendor authorized across your org and the data they reach.
See the engine in action.
Put these concepts to work across your company in minutes.