If an enterprise agent answers “What was revenue in Q1?” after the definition of revenue changes, valid SQL is not enough to make the answer reproducible. The agent also needs a clear rule for which definition to apply—and a record of the version it used. Versioning business semantics preserves the meaning behind answers as definitions evolve.
Why business definitions need versions
Suppose the earlier definition of revenue is “Recognized Revenue.” A later approved definition is “Recognized Revenue – Approved Adjustments.” Both can be calculated correctly, but they do not mean the same thing. If the newer definition silently replaces the older one, a later user may be unable to determine what an earlier answer meant.
As an Amazon Associate I earn from qualifying purchases.
Keep a stable identity for the business concept, such as revenue, and represent each materially different definition as a distinct version. Versioning the meaning alongside the SQL or other implementation lets an organization preserve prior interpretations without treating an old definition as current. This is a design recommendation, not a formal industry standard; the article’s guidance is to make semantic changes explicit and governed. The article’s discussion of semantic versioning
Recommended Free Tools
What to record for each semantic version
A semantic object should make its identity, meaning, status, timing, and implementation inspectable. A practical record can include:
#1 Best Overall
- Stable concept ID: an enduring identifier such as
revenue, independent of any one definition. - Version and definition: a version identifier and the human-readable rule, formula, or expression it represents.
- Owner and status: the accountable business or data owner and a lifecycle state such as draft, approved, published, or deprecated.
- Effective interval: the business dates for which the definition is intended to apply.
- Publication and approval provenance: when it was approved or made available, who authorized it, and any relevant rationale.
- Dependencies and physical mapping: related metrics, dimensions, source fields, tables, and transformations needed to understand downstream impact.
- Answer lineage: enough information to later reconstruct how an agent resolved the concept, including the selected version and mapping.
These fields help distinguish a change in business meaning from a change in implementation, and make it possible to inspect which downstream definitions or answers may be affected.
Keep publication time separate from effective time
A definition can be approved or published on one date but intended to apply to business activity from an earlier date. Those are different timelines. Record both the system-time event—when the organization accepted or published the version—and its business-effective interval. Otherwise, “the current definition” can be confused with “the definition that applied to the period being queried.”
For example, if a revised revenue rule is published in April but is intended to apply beginning January 1, the record should preserve both the April publication event and the January effective start. The specific dates in this example illustrate the distinction; they are not a claim about a particular company’s policy.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Choose how historical questions should be interpreted
Questions such as “What was Revenue in January?” are ambiguous after a definition change. The answer depends on whether the organization wants the meaning that applied at the time or the current rule applied to old data.
| Reporting choice | Meaning | Useful when |
|---|---|---|
| As was | Use the definition effective for the historical period, preserving the meaning used at that time. | The goal is to reproduce a historical report or understand what users were told under the then-current policy. |
| Restated | Apply the current definition to historical data. | The goal is to view prior periods under today’s definition. |
Neither choice is universally correct. The organization should define which policy applies to which reports and agent interactions, and make the chosen interpretation visible in the answer or its supporting lineage.
Set a policy for comparisons across versions
“Compare Q1 and Q3 Revenue” raises a separate issue: if a definition changed between the periods, the figures may not be semantically comparable even when both calculations succeed. An organization needs an explicit rule for such cases—for example, use each period’s effective definition, restate both periods under the current definition, or flag the comparison as spanning a change. The right choice depends on the purpose of the comparison; the important point is not to let an agent silently mix interpretations.
Govern changes before they become authoritative
A reliable rollout makes changes visible and reviewable rather than allowing a draft to become an agent’s default meaning. The article recommends a controlled lifecycle built around the following checks:
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problems- Classify materiality. Decide whether the proposed edit changes business interpretation or is only a non-semantic implementation adjustment.
- Review a semantic diff. Show how the new definition differs from the prior version in language and expression, so reviewers can assess the meaning change.
- Check dependencies. Identify affected metrics, reports, mappings, and agent-facing objects before publication.
- Validate the version. Confirm that its expression, source mappings, and intended effective interval are coherent and that dependent uses have been considered.
- Approve and publish deliberately. Preserve ownership and approval provenance, and keep draft or deprecated states from being treated as authoritative by default.
- Retain answer lineage. Log the resolved semantic object, version, effective date, and physical mapping used for each answer, so a later review can establish what produced it.
How current platforms can support shared business context
Platform features can provide building blocks for governed definitions, but their existence does not by itself establish that an organization has implemented the versioning and historical policies described above.
Databricks Unity Catalog semantics
Databricks documents Unity Catalog capabilities for organizing business terms and structures, reusable metric views, governed Pages, and trust signals such as certification and deprecation. Metric views separate measure definitions from dimensions and are documented for use across SQL, notebooks, dashboards, Genie Agents, alerts, and external BI. These are product capabilities, not evidence that a particular versioning policy or answer will be correct. See Unity Catalog business semantics and Databricks metric views.
Rank #4
Microsoft Fabric IQ
Microsoft describes Fabric IQ as a way to provide shared business context over OneLake data, Power BI semantic models, and ontology. Its ontology documentation covers entity types, properties, relationships, data bindings, and grounding for agents. Microsoft labels ontology as a preview feature, so its availability and status should be checked against current Microsoft documentation. These descriptions establish platform capabilities, not a guarantee that an agent will choose the intended historical definition. See Microsoft Fabric IQ and Fabric ontology overview.
What a reproducible answer should leave behind
When an agent returns a business metric, its trace should make it possible to answer: which semantic object did it resolve, which version did it use, which effective date or interval governed the choice, and which physical mapping produced the value? Capturing those details turns an answer from a number with hidden assumptions into an auditable result. Reproducibility requires preserving both the computation and the business meaning selected for it.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchQuick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
