Keystone EA™ is a complete, opinionated enterprise architecture standard — a layered metamodel, a rule-governed relationship system, six ready-to-use viewpoints, and a five-phase methodology to put it all into practice.
Most enterprise architecture frameworks are either too heavy to maintain or too loose to trust. Keystone EA™ is built for the middle: rigorous enough to keep a live repository consistent, light enough that a small team can actually run it.
It describes an organisation — its strategy, governance, business, data, applications and technology — as one model, where every element has a defined type and every link is governed by an explicit rule. From a board-level driver down to a database engine, everything traces through the same fabric.
The standard rests on three parts that fit together: a metamodel, a viewpoint library, and an implementation methodology.
Nineteen object types across seven layers, joined by six relationship types — with a complete rule set that defines exactly which links are valid between which objects.
Object types & rulesSix curated perspectives — governance, strategy, capability, process, application and technology — each showing only the object types that matter to a specific stakeholder question.
Browse viewpointsA five-phase implementation path — from Establish to Transform — across parallel Content and Architecture Readiness tracks, with AI-assisted workflows woven throughout.
See the phasesKeystone EA™ stacks the enterprise into seven layers. Upper layers express intent; lower layers deliver it. Every object type lives in exactly one layer.
Nineteen object types — not hundreds. Enough to model any enterprise faithfully, few enough that the model stays readable and a small team can maintain it.
A universal Connection means nothing is ever un-modelable, while a complete rules matrix governs the qualified relationships — so the model stays both flexible and consistent.
You never face the whole model at once. Viewpoints filter it down to the handful of object types that answer one stakeholder's question — governance, strategy, capability or integration.
The methodology assumes a live repository accelerated by AI — from structuring an AI service portal early to autonomous impact analysis and predictive insight at maturity.
Realization links thread a single line of sight from a strategic driver, through capabilities and applications, down to the infrastructure that ultimately delivers it.
Principles and standards are first-class objects, so governance and compliance are modelled — not bolted on — and the technology reference model is wired into governance.
Open the metamodel to inspect all nineteen object types and the interactive relationship rules matrix.