Glossary
Plain definitions of the terms that come up when a company governs, measures and pays for AI.
A
- AI audit trailA chronological record of the events, versions, decisions, approvals, and human actions needed to reconstruct how an AI system was used and governed.
- AI business caseEvidence used to decide whether an AI initiative should be funded, tested, expanded, redesigned, or stopped.
- AI cost optimizationThe practice of controlling the total cost of an AI system while preserving required quality, speed, reliability, and business outcomes.
- AI governanceThe system of roles, policies, decision rights, controls, and evidence used to direct and oversee AI throughout its lifecycle.
- AI measurement frameworkA structured method for deciding what to measure, where AI evidence comes from, and how an organization will judge an AI initiative.
- AI readiness assessmentAn assessment of whether a workflow, team and evidence system can measure the value of an AI investment.
- AI ROIA measure comparing the net financial benefit of an AI initiative with the total investment required to deliver it.
- AI spend managementThe practice of making AI costs visible, attributable, forecastable, and controllable across an organization.
B
- Baseline measurementThe process of recording how a defined workflow performs before an AI intervention so later outcomes can be compared fairly.
- Benefits realizationThe managed process of turning an expected improvement from an AI investment into an outcome the organization actually captures.
- Business value of AIThe measurable benefit an organization receives from using AI in a specific business context.
C
- Completed workA unit of business work that has reached a predefined acceptance state after required generation, review, correction, approval and system updates.
- Correction rateThe share of AI-influenced outputs that require a person to fix, revise, or replace them before accepted completion.
- Cost avoidanceA measurable future expense that an organization does not incur because an AI-enabled change reduces or removes the need for it.
- Cycle timeThe elapsed time from the start of active work on a defined item until accepted completion in an AI-enabled workflow.
D
E
F
- FinOps for AIA cross-functional operating practice for understanding, allocating, forecasting, and optimizing AI costs while connecting spend to business value.
- Fully loaded costThe complete cost of operating an AI-enabled workflow during a defined period, including technology, labor, implementation, review, governance, and support.
I
- Impact attributionThe process of estimating how much of an observed outcome occurred because of an AI intervention rather than other factors.
- Incremental valueThe additional business value produced because of an AI intervention beyond what would probably have occurred without it.
- Inference costThe cost of running a trained AI model to produce a prediction, classification, recommendation, or generated output.
- Initiative inventoryA governed record of the AI systems, tools, agents, pilots, and workflow changes an organization is evaluating or operating.
K
M
- Measurement windowThe defined period used to observe and compare cost, adoption, completed work, quality and outcomes for an AI-enabled workflow.
- Metric dictionaryA controlled record of the definitions used to measure AI investment, adoption, workflow performance, cost, quality, risk, and business outcomes.
O
- Observed valueA measurable improvement seen after an AI-enabled change before attribution or benefit capture has been fully established.
- Opportunity sizingThe process of estimating whether a defined workflow has enough addressable work and plausible improvement to justify further AI investment.
- Outcome metricMeasures of the operational or business result associated with an AI-enabled workflow.
P
- Payback periodThe time required for cumulative financial benefits from an AI investment to recover its investment cost.
- Pilot success criteriaThe pre-agreed conditions used to decide whether an AI initiative should scale, continue testing, change, pause, or stop.
- Productivity measurementEvaluation of whether an AI-enabled workflow produces more or better accepted output for the resources used while quality and risk remain within agreed limits.
- Proof of valueA bounded body of evidence showing whether an AI-enabled change produced a decision-relevant business outcome under defined operating conditions.
R
S
T
- Time to valueThe elapsed time between an agreed AI investment starting point and the first verified business outcome that meets a predefined value threshold.
- Token costThe charge associated with the token units an AI model processes and generates.
- Total cost of ownershipThe full lifecycle cost of acquiring, building, operating, governing, and retiring an AI system.
U
V
Put a policy on the AI tools
your teams use.
Bring a security rule, an agent boundary or a usage limit. See how it becomes a policy on your devices.
Windows macOS Linux