Use Cases
Application portfolio management is the use case where Sparx EA delivers its most immediate business value. When the EA repository is governed and connected via EA GraphLink, executives don’t ask architects for portfolio briefings: they query live dashboards or ask Copilot directly. The prerequisite is MDG governance: without consistent tagging, portfolio dashboards display noise instead of intelligence.
Key Takeaways
The reality in most organizations: an application inventory exists somewhere. It is almost certainly in a spreadsheet. It was probably accurate when it was created. It is now six months to two years out of date, maintained by one person when they have time, and consulted reluctantly by stakeholders who suspect the data isn’t reliable.
This isn’t a discipline problem. It’s an architecture problem. Spreadsheet-based portfolio management fails because the data is disconnected from the architectural context that gives it meaning. You can capture that an application is “Active” in a spreadsheet. You can’t easily capture which business capabilities it supports, what technology it runs on, what other applications it has dependencies with, and what the combined risk looks like if the platform it sits on approaches end of life.
Sparx EA changes this: if the repository is governed. That conditional matters. A populated but ungoverned repository doesn’t solve the problem; it moves it.
The specific problem EA-based APM solves: portfolio data that is connected, consistent, live, and accessible to the people who need it: not just to the architects who maintain it.
In an ArchiMate-based APM model, each application is an Application Component. The component carries all the metadata that makes portfolio management meaningful. Getting the element definition right: through MDG Technology: is the governance foundation.
Required tagged values for every Application Component in a managed portfolio:
These five fields are what make APM data useful. MDG Technology is what makes them mandatory: elements cannot be saved without them. Without MDG enforcement, some applications will have complete data and some won’t, which means dashboard output is partially trustworthy and partially not. Partially trustworthy portfolio data is worse than no dashboard at all; it creates false confidence.
Application Components in isolation tell you what applications exist. Application Components connected to business capabilities tell you what would break if an application were decommissioned.
For APM, each Application Component should carry relationships to:
These relationships are what enable the questions that matter: “What business capabilities are supported by applications rated low health?”: a question that combines criticality, health score, and business service relationships, and that can be answered by EA GraphLink if the relationships are in the model.
The technical mechanism: MDG Technology in Sparx EA allows you to define stereotypes that extend base ArchiMate elements. An “Application Component (Portfolio)” stereotype carries the five required tagged values as mandatory properties. When MDG is correctly defined and the Sparx EA environment is configured to enforce it, elements can’t be created without these properties.
This is the difference between a governance policy and a governance control. A style guide says “please fill in the lifecycle field.” MDG enforcement says the element doesn’t save without it.
EA GraphLink transforms the Sparx EA repository into an AI-accessible and BI-accessible data layer. For APM, this means portfolio data: all Application Components, their tagged values, and their relationships: becomes a live dataset that Power BI or Tableau connects to directly.
The operational outcome: when an architect updates an application’s health score from 4 to 2 in the repository, that change appears in the executive portfolio dashboard on the next refresh. No export. No manual update. No email to the analyst who maintains the spreadsheet.
A mature APM Power BI dashboard, connected via EA GraphLink, typically contains:
These dashboards are what architecture leaders bring to investment planning discussions. When the data is live and governed, the conversation changes: instead of defending the accuracy of a spreadsheet, architects are presenting decisions with live evidence.
EA GraphLink is not limited to the Microsoft ecosystem. The same live data connection works with Tableau for Salesforce-aligned organizations, and with other BI platforms that support standard data connection formats. The underlying mechanism is the same; the visualization layer adapts to what the organization already uses.
When EA GraphLink connects the repository to the Microsoft 365 ecosystem, Microsoft Copilot can answer portfolio questions using live architecture data. A business stakeholder in a Teams meeting can ask “which applications supporting the Finance capability are rated high criticality and low health?” and receive an answer synthesized from the current repository state: without involving an architect, without waiting for a report, without any training in how to read architecture diagrams.
This is the shift from architecture as a service to architecture as infrastructure. When the data is live and accessible, the architecture team’s value shifts from answering questions to improving the quality and depth of the data that answers them automatically.
Kernaro AI Hub is a purpose-built architecture intelligence platform that connects to the Sparx EA repository via EA GraphLink. For APM, it provides stakeholder-facing natural language query capabilities without requiring the Microsoft 365 integration path. A portfolio manager can browse to the Kernaro AI Hub interface, ask “what are the top five applications at end-of-life risk in the next 18 months?”, and receive a synthesized answer from live data.
Kernaro AI Hub is the right choice when the organization isn’t deeply invested in a single AI assistant ecosystem, or when architecture intelligence needs to be accessible to a broad stakeholder audience beyond the M365 user base.
EA GraphLink’s MCP (Model Context Protocol) server enables more complex analytical queries. “What would be affected if we decommissioned the legacy CRM platform?” is not a simple lookup: it requires traversing relationships from the CRM application to the business services that depend on it, to the data entities it manages, to the other applications that consume those data entities. EA GraphLink’s MCP layer handles this traversal and returns a synthesized impact analysis rather than a raw list of elements.
This is where most APM programs fail. Portfolio data is collected: often with genuine effort: without the MDG governance that enforces consistency. The result: a repository where some applications have complete lifecycle and health data, some have partial data, and some have none. The portfolio dashboard reflects this incompleteness, and stakeholders quickly learn not to trust it.
The Discover service assesses the current state of APM data quality in the repository. It answers the question: “Is the portfolio data in this repository trustworthy enough to connect to a live dashboard?” If the answer is no: and it often isn’t: the Deploy service establishes the MDG governance that makes portfolio data reliable. The Connect service then deploys the EA GraphLink integration once the data foundation supports it.
Starting with Connect before Discover or Deploy is possible. It’s also likely to produce a dashboard that accurately reflects poor data quality: which is worse than no dashboard, because it creates pressure to trust data that shouldn’t be trusted.
What is the best way to model an application portfolio in Sparx EA?
Model applications as ArchiMate Application Components with MDG-enforced stereotypes that carry required tagged values: lifecycle status (controlled vocabulary), business criticality (numeric 1-5), technical health score (numeric 1-5), application owner, and next review date. Connect Application Components to business capabilities using ArchiMate serving relationships and to technology infrastructure using realization relationships. The combination of element metadata and cross-domain relationships is what makes EA-based APM more powerful than a spreadsheet.
How do I connect Sparx EA portfolio data to Power BI?
EA GraphLink, deployed as part of a Connect engagement, creates a live data connection from the Sparx EA repository to Power BI. Portfolio data: Application Components, their tagged values, and their relationships: becomes a live Power BI dataset. When architects update the repository, the dashboard reflects the change on the next refresh. The Connect engagement covers EA GraphLink deployment, initial dashboard configuration, and governance of the data refresh process.
What tagged values should every application component have in an APM model?
At minimum: lifecycle status (Active / Planned / Retiring / Retired), business criticality (1-5), technical health score (1-5), application owner, and next review date. Organizations with more mature APM programs often add: vendor name, vendor support end date, annual cost, number of users, primary business domain, and CMDB identifier. Add tagged values that stakeholders will actually use: every field that isn’t used in a dashboard or decision is maintenance overhead.
How do I handle applications that span multiple business capabilities?
An Application Component can have serving relationships to multiple ArchiMate Business Functions or Capabilities: this is the correct way to represent cross-domain applications. In the portfolio dashboard, these applications appear in the context of each capability they support. For investment decisions, a high-criticality application that spans four business capabilities is a consolidation risk; EA GraphLink queries and Copilot can surface these multi-dependency relationships on demand.
Can I import data from a CMDB into Sparx EA for APM?
Yes. Sparx EA supports scripted imports via its API, and EA GraphLink can support bidirectional data flows with CMDB platforms. In practice, CMDB data quality is often the first obstacle: CMDBs frequently contain inaccurate or stale data that needs remediation before it improves the EA repository. A common pattern is to use the EA repository as the authoritative APM source and synchronize a curated subset to the CMDB, rather than importing CMDB data in bulk. The Discover service includes an assessment of CMDB-EA integration feasibility for organizations considering this path.
What MDG configuration is needed for APM in Sparx EA?
A minimum APM MDG configuration includes: a governed stereotype extending ArchiMate Application Component with required tagged values (as described above); controlled vocabulary tag definitions (not free-text) for lifecycle status; an MDG profile that prevents elements from being saved without required fields; and package-level MDG enforcement so that all elements created in the portfolio package inherit the correct stereotype. Organizations with multiple business domains should also define domain-specific sub-stereotypes or tagging conventions that enable domain-level filtering in dashboards.
How do I prevent portfolio data from going stale in the EA repository?
The primary mechanism is ownership assignment and governance cadence. Every Application Component has an assigned owner (captured as a tagged value); governance reviews flag elements whose “next review date” has passed. A Power BI governance dashboard: also accessible via EA GraphLink: can show the architecture team which elements are overdue for review. Kernaro Assist event agents can automate notifications when review dates pass. The governance cadence (typically quarterly for actively managed portfolios) needs to be an organizational process commitment, not just a technical configuration.
What does a Connect engagement cover for APM?
A Connect engagement for APM covers: EA GraphLink deployment and configuration for the organization’s repository structure; live Power BI or Tableau dashboard development using current portfolio data; Microsoft Copilot or Kernaro AI Hub integration for natural language portfolio queries; MCP server configuration for impact analysis queries; and documentation of the data model that the dashboards depend on. Connect assumes a governed repository: if MDG governance is not yet established, a Discover assessment and Deploy engagement should precede Connect.
Connect: EA GraphLink deployment and APM AI integration. $50K–$185K+ depending on integration scope and dashboard complexity.
Discover: If you’re not sure whether your current portfolio data quality supports a live dashboard, start here. MDG readiness assessment, data quality baseline, and roadmap. $25K–$75K.
Also relevant: Power BI Integration with Sparx EA | Kernaro AI Hub Overview
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.