A metric definition records how the business worked at one moment. Then you acquire a company, change the pricing model, or an engineer splits one status field into two, and the definition that was correct in March returns a wrong number in September. Nothing fails while this happens. No pipeline breaks, no error appears, and the reports keep rendering.
LazyFox watches every connected source for the changes that invalidate a definition, versions each approved change, and flags every query and report that depended on the old meaning.
Why governed meaning decays
Three ways definitions go wrong long after everyone agreed on them.
An engineer renames a column, changes a type, or splits one status into two. The query still resolves, so nothing announces itself. The metric built on top now measures something slightly different, and the only symptom is a number that still looks plausible.
The silent breakAfter an acquisition you have two customer records, two revenue recognition policies, and two readings of an active account. People settle it in meetings and spreadsheets. The semantic layer carries on serving the pre-acquisition answer to every AI tool that asks.
The definition gapA YAML file or a catalog entry records what somebody believed during the documentation sprint. Nothing in that file knows the business has changed, and nothing prompts a person to check. Definitions decay quietly because reviewing them is nobody's job.
The ownership voidHow LazyFox catches it
LazyFox monitors your connected systems for the schema and value changes that affect a governed definition: a renamed field, a new status code, a distribution that shifts after a policy change. It also watches how people phrase questions and what they correct in the answers. A term that suddenly needs qualifying every time somebody asks about it has usually moved.
Neither signal depends on anyone filing a ticket. Drift is found where it happens rather than reported after a wrong number reaches a board deck.
Explore the Semantic Manager →When a definition changes, LazyFox creates a new version rather than replacing the old one. The previous meaning stays queryable, so a figure in last quarter's reporting can still be traced to the definition that produced it, and an auditor asking why two quarters differ gets the governed answer instead of an archaeology project.
Domain experts propose the change in natural language and a semantic manager approves it. There is no YAML file to edit and no pull request to review.
See how answers stay defensible →A definition is rarely used once. It feeds dashboards, scheduled reports, downstream metrics, and whichever AI tools your teams have connected. When the meaning moves, LazyFox flags each of those dependencies rather than letting them keep answering from the old version.
The health map shows what is governed, what is contested, and what is currently out of date, so the semantic manager can see the state of your knowledge without auditing it by hand.
See how conflicts get resolved →What drift looks like
A governed definition of revenue, and the sequence that would have broken it quietly on any other stack.
Revenue is defined as Salesforce contracted ARR at close, approved by Finance, and published as version 1.0. Every AI tool and report now resolves the term the same way.
An SAP S/4HANA release changes the type of the orders amount column. The pipeline keeps running and no job fails. LazyFox flags the change against every definition that reads from that field.
Three dependent metrics and the reports built on them are surfaced with the conflict, before anyone presents a number derived from them.
Finance approves the corrected definition in natural language. LazyFox publishes version 1.1, keeps 1.0 queryable for prior reporting, and clears the flag on each dependency.
Semantic drift is what happens when a business metric keeps its name but changes its meaning. A definition of revenue, churn, or an active account is written to match how the business works at the time. Acquisitions, pricing changes, reorganisations, and upstream schema changes then move the underlying reality without touching the definition. Queries keep resolving and reports keep rendering, so the failure is invisible until somebody notices that two systems disagree or that a figure no longer matches last quarter.
LazyFox treats definitions as living records rather than static documentation. It monitors connected systems and query behaviour for the signals that a term has moved, versions every approved change so prior reporting stays traceable, and flags each dependent metric and report that was built on the previous meaning. Business users approve the corrected definition in natural language, without an engineering queue.
See it live
In the demo we will take a KPI whose definition moved after a pricing change, a reorganisation, or an acquisition, and show you how LazyFox surfaces the drift, versions the correction, and flags every report that was still answering from the old meaning.