For CIOs, VPs of Technology, and enterprise transformation directors.
Your architecture team is sitting on one of the most valuable data assets in your organization. The enterprise architecture repository: the structured record of your applications, business capabilities, technology components, data flows, and integrations: contains exactly the kind of information that executives, program directors, and AI tools need every day. Which applications support which business processes? Which technology components are approaching end of life? What are the dependencies that will be affected by this proposed change? Where are the integration points that a new system needs to connect to?
The problem is that most of this data is effectively locked away. Executives cannot access it without going through the architecture team. AI tools cannot query it. Business stakeholders cannot search it. The architecture team spends a substantial portion of its time fielding requests for information that already exists in the repository, reformatting it for the audience asking the question, and producing reports that are out of date almost as soon as they are published.
This is not a criticism of architecture teams: it is a structural problem with how architecture data is stored and exposed. Sparx Enterprise Architect and similar tools are powerful modeling and governance platforms, but they are designed for architects, not for the broader organization. The data they contain has historically been accessible only to people with licenses and skills to use the tool directly.
The result is a significant gap between the value the repository could deliver and the value it actually delivers. Architecture teams that could be focused on analysis, design, and governance spend their time on data extraction, formatting, and stakeholder communication. Executives who could be making faster, better-evidenced decisions wait for reports that may take days to produce. The architecture investment is real, but much of the potential return is unrealised.
This has begun to change.
Two developments in 2024 and 2025 have materially changed what is possible for enterprise architecture data.
The first is the GraphQL query interface. GraphQL is a well-established, standardized way of exposing structured data for query by external systems: BI tools, data integration platforms, and custom applications. EA GraphLink, a product from Sparx Systems released in this period, implements a GraphQL interface for Sparx EA repositories. This means Power BI, Tableau, and other BI platforms can query your architecture repository directly and display live, current architecture data in dashboards: without requiring data exports, manual updates, or architect intervention. A technology lifecycle dashboard that used to require quarterly exports and manual formatting can now refresh automatically from the live repository.
The second development is the Model Context Protocol (MCP). MCP is an open standard: now supported by Microsoft, Anthropic, and a rapidly growing ecosystem of AI tool vendors: for giving AI assistants access to structured external data as part of a conversation. When an AI assistant supports MCP, it can query a connected data source in real time and use what it finds to answer questions accurately, rather than relying solely on what it was trained on.
EA GraphLink’s second interface implements MCP for Sparx EA. This means that any MCP-compatible AI assistant: including Microsoft Copilot, Claude, and others: can be connected to your architecture repository. An architect asking Copilot “which applications share an integration with the Payments Platform?” gets an answer drawn from live repository data. An executive asking “what is the technology lifecycle status of our core banking applications?” gets a current, accurate response: not a hallucinated one.
These two changes together: live GraphQL for BI, live MCP for AI: are what makes AI augmented architecture practically achievable in 2025 and 2026. The tools exist, the standards are stable, and the implementation path is clear. What varies between organizations is how ready the repository is to support it.
Three specific changes become possible once the connection is made between your architecture repository and the AI and BI tools your organization already uses.
1. Architects spend less time on reporting and more time on architecture.
A senior enterprise architect in a typical large organization spends between 30% and 50% of their time producing outputs that consumers cannot get any other way: portfolio summaries, impact assessments, technology inventories, and ad-hoc responses to stakeholder requests. Once architecture data is accessible through BI dashboards and AI assistants, a significant portion of those requests are answered without architect involvement. The architect team does not shrink: it redirects. The time recovered goes to analysis, design work, and governance activities that genuinely require architectural expertise.
This is not a theoretical improvement. Organizations that have made their architecture data live and queryable consistently report that their architecture teams become more strategic: not because the team changes, but because the information bottleneck is removed.
2. Executives can ask direct questions about the application portfolio and get answers grounded in live data.
The most common complaint from executives about enterprise architecture is that they cannot get quick answers. They understand that the architecture team has the information; they do not understand why getting it requires a multi-day process. An AI assistant connected to the architecture repository changes this fundamentally.
A CTO preparing for a board conversation about technology debt can ask Copilot directly and get a structured summary of applications approaching end of life, the business capabilities they support, and the migration status of any active remediation programs: all from current repository data. A VP of Technology evaluating a vendor proposal can ask which systems would be affected by a change to a shared integration platform and get a dependency map in seconds rather than days.
The answers are only as good as the repository data: which is why repository quality and MDG governance are prerequisites, not afterthoughts. But when the repository is well-governed, this kind of executive self-service is genuinely achievable.
3. Governance decisions are faster because the evidence is always current.
Architecture Review Boards and technology investment committees typically operate from point-in-time data: the architecture team produces an analysis, the committee reviews it, and by the time a decision is made, the analysis may be weeks old. When architecture data is live and queryable, the evidence base for governance decisions is always current.
This changes the cadence and confidence of governance. Committees can ask questions in the meeting and get answers from current data. Dependency analysis for a proposed project change does not require a two-week preparation cycle. The architecture team presents less and advises more.
EA GraphLink is a product from Sparx Systems that provides the connectivity layer between Sparx Enterprise Architect and external systems. It exposes two interfaces: a GraphQL interface (Interface A) for BI tools and data integration platforms, and an MCP interface (Interface B) for AI assistants. It requires a Sparx EA deployment running on a shared database repository (SQL Server, Oracle, or PostgreSQL).
Kernaro AI Hub is a stakeholder-facing platform, reaching general availability in 2026, that provides a purpose-designed interface for business and technology leaders to explore EA insights. It requires EA GraphLink as its connectivity layer. It is aimed at audiences who will not use Sparx EA directly but need current, reliable architecture insight to do their jobs.
Both products are made by Sparx Systems and purchased directly from Sparx Systems. Sparx Services does not resell software licenses. Our role in a program that includes these products is to scope the license requirements, provide the bill of materials, and deliver the implementation: covering EA GraphLink configuration, integration design, repository preparation, and downstream BI and AI tool connection.
MDG Technology is Sparx EA’s built-in metamodel governance framework. It is not a separate product: it is the configuration layer within Sparx EA that defines stereotypes, tagged values, relationship rules, and validation checks. It is the primary quality mechanism that makes AI queries reliable. An organization without mature MDG governance will find that AI tools connected to their repository produce inconsistent or unreliable results.
Sparx Services structures AI augmentation engagements in three phases, which can be run sequentially or partially in parallel depending on the organization’s starting point.
Phase 1: Discover ($25,000 – $75,000 | 2–6 weeks)
Discover is a structured assessment of your current EA capability and platform health. It covers a repository audit (quality, completeness, MDG maturity), stakeholder interviews (with architects, business stakeholders, and executive consumers), platform review (deployment configuration, shared repository health, version currency), and an AI readiness assessment scored against the five dimensions in the Sparx Services AI Readiness Framework.
The output is a prioritized findings report and recommended program roadmap: specifying what needs to be done, in what sequence, and with what investment, to achieve AI augmented architecture capability. Discover is the right starting point for any organization that does not have a clear picture of where its gaps are.
Phase 2: Amplify ($45,000 – $160,000 | 4–16 weeks) and/or Deploy ($30,000 – $130,000 | 3–12 weeks)
Amplify addresses repository governance: MDG Technology configuration, naming standards, tagged value disciplines, element quality uplift, and team enablement. It is the work that makes the repository AI-ready. Deploy addresses platform deployment or re-deployment: the technical infrastructure that underpins the repository.
For organizations with a mature, well-configured platform and good governance, one or both of these phases may be abbreviated or skipped. For most organizations, some Amplify work is needed before Connect will deliver full value.
Phase 3: Connect ($50,000 – $185,000+ | 6–20 weeks)
Connect is the AI integration program. It covers EA GraphLink deployment and configuration, GraphQL interface setup and BI tool connection (Power BI, Tableau), MCP interface setup and AI tool connection (Copilot, Kernaro AI Hub), stakeholder enablement, and the measurement framework that shows the before/after impact.
The total program timeline for an organization starting from Discover through to a live AI integration is typically 6 to 12 months, depending on starting maturity and scope. Investment across the three phases ranges from approximately $120,000 to $420,000+, depending on scale, complexity, and the scope of integration surfaces. Software license costs (EA GraphLink, Kernaro AI Hub) are in addition to services and are purchased directly from Sparx Systems.
The following five questions give you a practical read on your organization’s readiness for AI augmented architecture. They are framed as questions your board or audit committee might ask: because in increasingly technology-dependent organizations, they often do.
Sparx Services offers a free 20-minute discovery conversation for CIOs, VPs of Technology, and transformation directors exploring AI augmented architecture. The conversation covers your current EA maturity, the questions your organization needs architecture to answer, and whether a Discover engagement or a direct Connect scoping discussion is the more appropriate starting point.
There is no obligation and no sales presentation. The conversation is structured and practical: useful regardless of whether you engage Sparx Services.
Book at sparxservices.com/contact
Sparx Services: Enterprise Architecture Platform Specialists
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.