Use Cases
Architecture data locked in a modeling tool helps architects. Architecture data accessible to the people making business decisions changes how organizations work. This page describes the options for getting EA repository intelligence to business stakeholders: from live dashboards to natural language AI queries. None of these paths require an EA license or training in architecture notation.
Key Takeaways
You’ve been told the organization has an enterprise architecture team. You’ve occasionally received a PowerPoint presentation with architecture diagrams. The diagrams were dense with notation you weren’t trained to read. The presentation was three months old by the time it reached you. When you asked a follow-up question about which systems support a specific business process, you were told the architect is busy and would send something next week.
This is the normal state of architecture visibility in most organizations: not because architects don’t want to be helpful, but because the tools they use to do their work aren’t built for executive consumption. A Sparx EA diagram is designed for modeling precision. It is not designed for a business leader who needs a 30-second answer to a portfolio question in the middle of a budget conversation.
The problem has two sides. The architecture team spends significant time translating their repository data into stakeholder-readable formats: reports, presentations, dashboards: instead of doing architecture work. And stakeholders still don’t get real-time answers; they get the translation that was completed last week.
EA GraphLink, and the stakeholder access paths it enables, addresses both sides of this problem simultaneously.
When the EA repository is connected via EA GraphLink, architecture data becomes accessible in the tools and formats stakeholders already use. The architecture team doesn’t need to produce translations. Stakeholders don’t need to wait for them. Four access paths serve different stakeholder contexts:
Live portfolio dashboards delivered in the BI platform the organization already uses. Application health, lifecycle status, compliance coverage, business capability support maps: updated automatically when architects update the repository.
The stakeholder experience: the same reporting environment they use for financial data and operational metrics now includes architecture intelligence. No new tool. No training. The portfolio health quadrant is a tab in the same Power BI workspace as the monthly operational review.
When Power BI or Tableau is the right path: the organization has an established BI platform, stakeholders already use it for other reporting, and the primary need is portfolio and health dashboards rather than conversational queries. This is the lowest adoption friction of any access path.
Natural language architecture queries through the AI assistant the organization already uses. A business stakeholder in a Teams meeting asks “what applications are at risk in the next 18 months?” and receives a synthesized answer from live repository data: no architect involvement, no report request, no waiting.
When Copilot is the right path: the organization is invested in Microsoft 365 and Microsoft Copilot is already being deployed or is in use. Architecture intelligence becomes one of the data sources Copilot draws from, alongside email, documents, and calendar data. For Salesforce-aligned organizations, the same principle applies via Salesforce Agentforce.
The experience for the business leader: architecture data answers arrive in the same interface where they ask questions about anything else. The boundary between “architecture questions” and “business questions” disappears, because the data is accessible in the same place.
A purpose-built architecture intelligence platform for stakeholders and architects. Browser-based, no EA license required, no installation. Optimized specifically for architecture-domain queries rather than general-purpose AI assistant use.
When Kernaro AI Hub is the right path: the organization isn’t heavily committed to a single AI assistant ecosystem, or wants architecture intelligence available to stakeholders who aren’t in the M365 or Salesforce ecosystem. Kernaro AI Hub provides a dedicated, governed interface for architecture queries: with controls over what data is accessible and to whom.
The experience: a browser-based interface where stakeholders type questions about the organization’s architecture and receive answers synthesized from live repository data. Questions can range from portfolio health (“which applications are high criticality and low health?”) to dependency analysis (“what systems would be affected by a migration of the customer data platform?”) to compliance coverage (“which business processes have no documented architecture?”).
Browser-based access to structured EA repository views: diagrams, element lists, architecture views organized by domain or business capability. Prolaborate doesn’t require the full Sparx EA client and is lower-cost for broad stakeholder rollouts.
When Prolaborate is the right path: stakeholders need to browse and explore architecture views rather than ask questions. The primary use case is making architecture documentation accessible to a large audience: project managers, business analysts, solution architects who need read access to the repository without the modeling capabilities of the full client. Lower cost than AI integrations; more appropriate for document-centric access patterns.
| Stakeholder Need | Best Path |
|---|---|
| Portfolio dashboards in existing BI environment | Power BI or Tableau |
| Natural language queries in Microsoft 365 | Microsoft Copilot |
| Natural language queries in Salesforce ecosystem | Salesforce Agentforce |
| Dedicated architecture AI, ecosystem-independent | Kernaro AI Hub |
| Browse architecture views and diagrams | Prolaborate |
| Mix of dashboards and queries, Microsoft-aligned | Power BI + Copilot |
In practice, most mature EA stakeholder programs combine two paths: a BI dashboard for regular portfolio reporting, and a natural language AI tool for ad-hoc queries. These serve different stakeholder interaction patterns and complement each other.
Plain language: if the architecture data in the repository is inconsistent or incomplete, none of these tools help. They amplify what’s there.
A Copilot query about application lifecycle status returns correct answers when every application has a governed lifecycle tag. It returns partial answers when 60% of applications have lifecycle tags. It returns no useful answer when the lifecycle field is populated with inconsistent free-text values that mean different things to different architects.
This is the MDG governance prerequisite. Before stakeholder access paths are deployed, the architecture data they draw from needs to meet a quality threshold: consistent tagging, complete required fields, governed relationship structure. The Deploy service establishes this foundation. The Discover service assesses whether the current repository is ready.
The architecture team doesn’t need a perfect repository before deploying stakeholder access. They need a repository where the data that stakeholders will query is consistent and trustworthy. Defining the scope of that data and ensuring its quality is what makes stakeholder access useful rather than counterproductive.
How do I get architecture data into my Power BI reports?
EA GraphLink creates a live data connection from the Sparx EA repository to Power BI. Portfolio data: applications, their properties, their relationships to business capabilities and technology infrastructure: becomes a Power BI dataset that refreshes automatically when the repository is updated. A Connect engagement deploys EA GraphLink, designs the Power BI data model, and builds the initial dashboard set for your portfolio and stakeholder reporting needs.
Can I ask architecture questions in Microsoft Teams without contacting an architect?
Yes, when Microsoft Copilot is integrated with EA GraphLink as a data source. A business stakeholder in Teams can ask portfolio, dependency, or risk questions and receive answers synthesized from live repository data. The quality of those answers depends on the repository data quality: well-governed repositories produce accurate Copilot responses; ungoverned repositories produce partial or inaccurate ones.
What is Prolaborate and how is it different from accessing the full Sparx EA tool?
Prolaborate is a browser-based stakeholder portal for Sparx EA repositories. It provides read access to architecture views, diagrams, and element information without the modeling capabilities of the full EA client. Prolaborate is appropriate for stakeholders who need to browse and reference architecture documentation: it is not an AI query tool. Cost per user is lower than a full EA license, making it viable for broad organizational rollouts. It is developed by Prolaborate (a separate company) and integrates with Sparx EA repositories via the standard Sparx EA API.
What is Kernaro AI Hub and who is it designed for?
Kernaro AI Hub is a purpose-built architecture intelligence platform that connects to Sparx EA via EA GraphLink. It is designed for architecture teams and their stakeholders who want natural language query access to architecture data without being embedded in a specific AI ecosystem (Microsoft, Salesforce, etc.). It provides governed, architecture-domain-optimized query capabilities with controls over data scope and access. It is appropriate for organizations that want architecture intelligence as a standalone capability rather than an extension of an existing AI assistant.
Do I need an EA license to use Kernaro AI Hub or Prolaborate?
No. Both tools access the EA repository through APIs rather than through the EA client. Stakeholders using Kernaro AI Hub or Prolaborate do not require an EA license. This is a significant cost and adoption consideration for organizations deploying stakeholder access at scale: it makes broad access economically viable in a way that full EA license rollouts are not.
What kind of questions can stakeholders ask the architecture AI?
Questions that the repository data can answer. Portfolio questions: “which applications are rated high criticality and low health?” Dependency questions: “what systems support the customer onboarding process?” Risk questions: “which business capabilities have no redundant application support?” Lifecycle questions: “what technology becomes unsupported in the next 18 months?” The scope of answerable questions is defined by the data in the repository: well-governed repositories with rich relationship models can answer richer questions.
How accurate are AI answers about our architecture?
Accuracy is a function of repository data quality, not AI capability. When tagged values are consistently populated and relationship models are complete, AI answers are accurate. When they are not, AI answers are confidently wrong: which is worse than incomplete. This is why MDG governance precedes stakeholder AI access in the recommended service sequence. The Discover service includes a data quality assessment that identifies the current accuracy floor for AI-answered queries.
What needs to happen before we can give business stakeholders AI access to architecture data?
Three things: repository governance needs to meet a quality threshold for the data scope stakeholders will query; EA GraphLink needs to be deployed and configured for the target access path; and the stakeholder access tool (Copilot, Kernaro AI Hub, or other) needs to be deployed and integrated. In practice, repository governance is the longest lead-time item. The Connect service deploys the integration; the Deploy service (typically preceding Connect) establishes the governance foundation.
Connect: EA GraphLink deployment and stakeholder access path integration. $50K–$185K+ depending on the number of access paths and integration complexity.
Discover: If you’re not sure which access path fits your organization’s stakeholder needs, or whether the repository is ready, start here. $25K–$75K.
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.