Resources
A non-technical guide for VPs, Directors, and Program Sponsors
Your organization’s IT landscape is complex. Behind every business capability: processing a customer order, issuing an insurance policy, delivering a patient prescription: sits a network of technology systems that make it work. Some of those systems are modern and well-maintained. Some are approaching end of life. Some have not been updated in a decade and are held together by integrations that nobody fully understands.
Decisions about which systems to invest in, which to replace, and how they connect are made by architects. Your architecture team maintains the authoritative picture of what technology exists and how it fits together. They know which systems support which business functions, which technology investments are on track, and which legacy platforms represent risk to your operations.
The problem is access. Architecture knowledge typically lives in a specialist tool: Sparx Enterprise Architect, ARIS, or similar: that business leaders cannot open, cannot query, and rarely see. The result is a persistent gap between the people who need architecture intelligence and the people who have it.
That gap has real consequences:
Decisions made on incomplete information. When a program board asks “is our customer data platform ready to support the new product launch?”, the answer typically comes back as a PowerPoint slide prepared by an architect: not a live answer drawn from authoritative data. The slide is accurate as of when it was prepared, which may have been two weeks ago. The actual data may have changed.
Expensive surprises at program go-live. The integration dependencies that matter most are often the ones nobody documented clearly. A major system replacement that discovers undocumented integration problems at go-live: after the new system is already live in production: is a program failure that happens more often than it should, and more often than it needs to.
Architecture insights that do not reach the board. The architecture team produces excellent analysis. The analysis sits in a tool that board members cannot access. A summary reaches the board as a quarterly report. By the time strategic decisions respond to that analysis, it may be six months old.
The underlying problem is structural: architecture tools are designed for architects, and there has been no practical way to make their content accessible to business leaders without a human intermediary translating and presenting.
That has now changed.
In the past 18 months, a new category of technology has emerged that connects architecture tools directly to the AI assistants that business leaders are already using.
You may already use Microsoft Copilot in your daily work: asking it to summarize a document, draft an email, or retrieve information from your organization’s Microsoft 365 environment. The same Copilot that does those tasks can now be connected to your organization’s architecture repository.
The technology that makes this connection possible is called EA GraphLink: a product made by Sparx Systems, the company behind the most widely used enterprise architecture tool in the world. EA GraphLink creates a bridge between the architecture repository and AI assistants like Microsoft Copilot, Salesforce Agentforce, and others.
Once that bridge is in place, a business leader can ask their AI assistant a direct question about the architecture:
“Which systems support our customer onboarding process?”
“What technology investments in the finance division are flagged as at risk?”
“Are there any applications approaching end of life that support our regulatory reporting function?”
The AI assistant retrieves the answer directly from the architecture repository: not from a document someone prepared last month, but from the live, maintained model that the architecture team updates continuously.
This is not a future state or a prototype. Organizations are using this capability in production today. The technology is mature. The constraint is not the AI: it is whether your architecture repository is well-maintained enough to produce reliable answers. (We will come back to this.)
Three specific capabilities become available to business leaders when AI-augmented architecture is in place. None of them require technical training. None of them require access to an architecture tool.
1. Ask your AI assistant direct questions about which systems support your business area
In Microsoft Copilot: or whichever AI assistant your organization uses: you can ask questions about your technology landscape in plain English and receive structured, accurate answers drawn from the live architecture repository.
This sounds simple, but it represents a significant change from the current state. Today, if you want to know which systems support your operations in a particular business unit, you either ask your architecture team (and wait for a response) or accept that you do not know. With AI-augmented architecture, you ask Copilot and receive an answer in seconds: the same answer your architect would give you, drawn from the same data source.
The practical applications are wide. You can ask about your own business area before a board meeting to verify that the technology picture you are presenting is accurate. You can ask whether a specific system your team relies on is flagged as a technology risk. You can ask what the integration dependencies are for a system your program is planning to replace. You ask; Copilot answers; you have the information.
2. Get live answers about which technology investments are on track or at risk
Architecture teams track the status of technology programs: which initiatives are on plan, which are behind, which have identified risks, which are blocked. This information currently reaches business leaders through program status reports, risk registers, and periodic briefings. By the time it arrives, it reflects the situation as of the last report cycle.
With AI-augmented architecture, the program status is in the live repository and accessible on demand. You can ask Copilot on the morning of a board meeting: “What is the current status of the CRM replacement program?” and receive an answer that reflects the repository state as of that morning: not as of last month’s report.
This changes how you prepare for and participate in governance conversations. You arrive informed, with access to the same source of truth that your program managers are working from.
3. See portfolio heat maps that update automatically
One of the most valuable outputs of an architecture practice is the portfolio heat map: a visual summary of the technology estate showing which systems are healthy, which are aging, which carry high business risk, and which are scheduled for replacement. Portfolio heat maps help boards and directors make technology investment decisions with a clear picture of the landscape.
Today, portfolio heat maps are typically produced quarterly by the architecture team and presented as static slides. They are accurate at the time they are produced and outdated by the time they are presented.
With EA GraphLink, portfolio heat maps are Power BI dashboards connected directly to the live architecture repository. When the architecture team updates the status of a system: because a vendor has announced an end-of-support date, or a risk assessment has been completed, or an investment decision has been approved: the dashboard reflects the change immediately. You see the current picture, not last quarter’s picture.
Program sponsors and directors with access to the Power BI dashboard can monitor their relevant portfolio area in real time, without waiting for the architecture team’s next reporting cycle.
AI-augmented architecture does not replace the architecture team. It amplifies them: making their work accessible to a wider audience and making their analysis available on demand. But the quality of the AI outputs is determined entirely by the quality of the data the architecture team maintains.
This is the critical dependency that every business leader considering AI-augmented architecture needs to understand: the AI is only as good as the architecture repository it queries.
An architecture repository that is well-maintained: with systems documented accurately, ownership recorded, integration dependencies mapped, technology risk assessments current: produces AI responses that are genuinely useful and reliable. A repository that is partially maintained, with missing data, outdated status, and uncaptured dependencies, produces responses that are incomplete or misleading. The AI reports what is in the repository; it cannot invent what is missing.
This means AI-augmented architecture requires two things from your architecture team:
A well-governed architecture repository. The architecture team must maintain the repository consistently: updating system status when things change, capturing new systems as they are procured, recording integration dependencies as they are understood. This is not new work: good architecture teams already do this. But it must be done with discipline, because the quality of business leader answers depends on it.
EA GraphLink: the connectivity layer. EA GraphLink is the technical product that connects the architecture repository to AI assistants and Power BI. It requires installation and configuration by someone with Sparx EA expertise. This is a one-time implementation effort, not ongoing maintenance. Once EA GraphLink is in place, the connection between the repository and your AI assistants is live and self-updating.
The business leader’s role is to create the expectation that these two things will be in place: and to sponsor the investment that makes them possible. If your architecture team does not have a well-governed repository, that is the first investment. If EA GraphLink is not yet deployed, that is the second.
Five diagnostic questions that will tell you where your organization stands on AI-augmented architecture:
1. “Do we have a single, maintained architecture repository that all architects work from?”
The answer reveals whether your architecture knowledge is consolidated or fragmented. Fragmented knowledge: spreadsheets, SharePoint documents, individual Visio diagrams: cannot be connected to AI tools. A single governed repository is the prerequisite. If the answer is “no” or “sort of”, the first investment is consolidation.
2. “Is our application inventory complete and current?”
An application inventory that is 12 months out of date, or that covers only some business units, will produce incomplete AI responses. Ask specifically: how many applications are in the inventory, when was it last verified, and which business units are covered? The answer tells you whether the AI queries on system ownership and risk will be reliable for your specific business area.
3. “Are our integration dependencies documented?”
Integration dependencies: which systems connect to which: are the most critical architecture knowledge for program risk management. If the architecture team cannot tell you which systems would be affected by replacing a core platform, the integration dependencies are not documented. This is the most common gap, and the one most likely to produce expensive go-live surprises.
4. “Do we have EA GraphLink deployed?”
EA GraphLink is the specific connectivity product that enables AI augmentation of Sparx EA repositories. If the answer is no: and most organizations have not yet deployed it: the AI-augmented architecture capability described in this guide is not yet available to you. The deployment is a defined implementation engagement, typically 8–16 weeks.
5. “Which AI assistants could we connect to our architecture repository?”
The answer depends on what AI tools your organization already uses. Microsoft Copilot, Salesforce Agentforce, Azure OpenAI, and others all support the connection protocol that EA GraphLink uses. If your organization already has Microsoft 365 with Copilot, the connection is straightforward. Understanding which AI tools are in scope shapes the implementation approach and timescale.
Sparx Services is an enterprise architecture consulting firm specializing in Sparx EA implementation, EA GraphLink deployment, and the governance frameworks that make AI-augmented architecture reliable. Our Connect engagement delivers the capability described in this guide.
What a Connect engagement delivers:
What Connect costs:
Connect engagements start at $50,000 and scale with the scope of the Power BI dashboard suite and the number of AI assistant integrations. A typical engagement for a mid-size organization runs 10–16 weeks. The investment does not include Microsoft Copilot or Power BI licensing, which is licensed separately through Microsoft.
If your architecture repository needs to be established or improved before Connect, a Discover engagement ($25K–$75K) assesses the current state and designs the repository governance model. A Deploy engagement ($30K–$130K) then builds the repository foundation before Connect delivers the AI layer.
What the outcome looks like for you:
Six months after a Connect engagement, a business leader in your organization will be able to ask Microsoft Copilot a question about their technology landscape and receive a reliable answer in seconds. Your program board will have a live portfolio dashboard they can consult before board meetings: not a quarterly slide deck. Your architecture team will spend less time answering stakeholder queries and more time doing the architectural analysis that only they can do.
If you have questions about AI-augmented architecture: or if you want to understand where your organization stands relative to what is described in this guide: contact the Sparx Services team. We can give you an honest assessment of what your organization needs and what it would cost, without a sales process.
Read more about Connect engagements
Read more about EA GraphLink and how it works
Sparx Services is an authorized Sparx Systems partner specializing in enterprise architecture implementation and AI-augmented architecture for organizations across financial services, government, energy, healthcare, aerospace, and defense.
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.