Two Mints, One Currency

Opening Scene

A large kingdom might operate several mints across its territory, for practical reasons of geography and logistics, but every one of them strikes coins to the exact same certified standard, coordinated centrally so a coin from one mint is fully interchangeable with a coin from another. Two mints, many physical locations, but genuinely one single currency, not two competing ones that happen to look similar.

Federated semantic layers across a large organization’s different teams or domains need this exact same coordination, and getting the balance right is a genuinely hard organizational problem.

In Plain English

A federated semantic layer allows different teams or domains to manage their own specific metric definitions somewhat independently — reflecting real, legitimate differences in local expertise and need — while still adhering to shared, organization-wide standards for how metrics are named, structured, and governed, so that metrics from different domains remain consistent and interoperable where it actually matters.

The Old Way

Organizations often faced a stark, unsatisfying choice: full centralization, where every metric definition required approval from one central team, creating a bottleneck as the organization scaled, or full decentralization, where every team defined its own metrics with essentially no coordination, reintroducing exactly the sprawl problem covered in Article 4 at an organizational scale.

Neither extreme worked well in practice. Full centralization couldn’t keep pace with a large, fast-moving organization’s genuine need for domain-specific metric expertise. Full decentralization sacrificed the consistency that made a semantic layer valuable in the first place. Organizations that hadn’t found a workable middle ground tended to oscillate uncomfortably between the two, dissatisfied with both.

What’s Changing (and Why AI Is the Reason)

Modern semantic layer platforms increasingly support genuine federation as a first-class architectural pattern. Rather than an all-or-nothing choice, mature platforms let domain teams own their specific metric definitions within a shared framework of naming conventions, data types, and governance standards enforced centrally, echoing the federated governance themes covered in this site’s data-fabric-mesh topic.
AI-assisted standards enforcement can maintain consistency across federated domains without requiring centralized manual review of every definition. Rather than a bottlenecked central team reviewing everything, AI-assisted tooling can automatically check that domain-specific metric definitions comply with shared organizational standards, catching violations without requiring a human reviewer in the loop for every single change.
AI-assisted cross-domain metric composition can help combine metrics from different federated domains correctly. When a metric genuinely needs to draw on definitions from multiple domains — combining a sales metric with a support metric, for instance — AI-assisted analysis can help verify the composition is semantically valid, catching the kind of subtle cross-domain inconsistency that’s easy to introduce accidentally.

The Metaphor, Fully Extended

Mint Element	Federated Semantic Layer Concept
Several regional mints, each striking coins to one shared certified standard	Multiple domain teams, each managing metrics within one shared organizational standard
One central mint trying to strike every coin for an entire kingdom	Full centralization, with one team bottlenecking every metric definition
Many uncoordinated regional mints each inventing their own currency	Full decentralization, with every team defining metrics independently and inconsistently
Royal inspectors verifying every regional mint’s output meets the certified standard	AI-assisted standards enforcement automatically checking domain metrics against shared standards
An exchange office correctly converting between coins from two different regional mints	AI-assisted cross-domain metric composition correctly combining metrics from different domains

For Beginners: What to Actually Do

Practice recognizing federation as a genuine middle ground between full centralization and full decentralization, not simply a compromise that satisfies neither goal.
Get comfortable with the idea that domain teams having some real autonomy over their own specific metrics is legitimate and valuable, not a governance failure to be eliminated.
Notice what shared standards actually need to be enforced centrally (naming conventions, core definitions, governance process) versus what can reasonably be left to domain-specific judgment.
When combining metrics from different domains, practice explicitly verifying the combination is semantically valid, rather than assuming compatibility just because both metrics technically exist.

For Practitioners and Leaders: The Deeper Layer

Evaluate whether your organization’s current semantic layer approach is genuinely federated, or whether it’s actually oscillating uncomfortably between full centralization and full decentralization without a coherent middle ground.
Adopt semantic layer platforms and architectural patterns that support genuine federation, allowing domain ownership within a shared standards framework rather than forcing an all-or-nothing choice.
Use AI-assisted standards enforcement to maintain consistency across federated domains without recreating the central bottleneck full centralization originally caused.
Use AI-assisted cross-domain composition checking specifically for metrics that combine definitions across domain boundaries, where subtle inconsistency is easiest to introduce accidentally.

Quick Recap

A federated semantic layer allows domain teams real autonomy over their specific metric definitions while adhering to shared, organization-wide standards for naming, structure, and governance.
Organizations without a genuine federated approach often oscillated between full centralization (a bottleneck) and full decentralization (a return to sprawl), satisfied with neither.
Modern semantic layer platforms increasingly support genuine federation as a first-class pattern, and AI-assisted standards enforcement can maintain consistency without a central review bottleneck.
AI-assisted cross-domain composition checking helps verify that metrics combined across federated domain boundaries remain semantically valid.

Where This Fits in the Series

Article 10 covered launching a new metric domain coherently. This article covered coordinating many domains under one shared standard. Article 12 looks at what happens when a metric’s meaning genuinely depends on context.