Analyzing and Controlling Architectural Debt with AI
- Mark Miller
- 6 days ago
- 2 min read
From Nadzeya Stalbouskaya at The Artificial UnIntelligence Conference: Most organizations operate with an invisible burden of architectural debt. It accumulates quietly through temporary fixes, disconnected changes, and undocumented dependencies. In this talk, I’ll share how we began to systematically identify and control architectural debt and how implementing an AI agent helped us not only track it but turn it into a decision-making tool.
Session Overview: Controlling Architectural Debt
This session is based on real-world experience: from architectural chaos to intentional, sustainable design.
1. Introduction: What is architectural debt and why is it more dangerous than technical debt Key distinctions between technical debt and architectural debt Real-world examples where “everything works” — but it’s too expensive, fragile, or hard to change Why architectural debt is often invisible — until it’s too late
2. How architectural debt accumulates in organizations Types of debt: conceptual, integration, process, structural Root causes: lack of architectural oversight, deadline pressure, chaotic automation Where to look: systems with unclear ownership, legacy integrations, redundant functionality
3. The AI agent as architect: how AI helps us identify, categorize, and manage debt Practical case: how we deployed an AI agent to collect metrics and surface architecture issues What the agent does: mapping dependencies, identifying outdated components, spotting violations Visualizing architectural debt and converting it into an actionable roadmap
4. What controlling architectural debt enables Increased transparency, faster change delivery, reduced operational risk Building a living architectural model instead of static diagrams Shifting from reactive fire-fighting to proactive architecture
Conclusion: Architecture is not a diagram — it's a long-term responsibility Debt control as a sign of architectural maturity Practical practices to implement tomorrow How to measure and communicate architectural debt to the business
Nadzeya's session, "Analyzing and Controlling Architectural Debt with AI", will be streamed globally September 16-17, 2025 as part of The Artificial UnIntelligence Conference.
About Nadzeya Stalbouskaya
Nadzeya Stalbouskaya is a Technology Lead and Enterprise Architect at IAG GBS, where she leads digital transformation and technology strategy initiatives across complex enterprise environments. With certifications from The Open Group in TOGAF and recognition as a member of the ICMG Enterprise Strategy & Architecture Advisory Group, she is passionate about redefining how architecture supports business success. Nadzeya specializes in building scalable, secure systems while mentoring women in tech and advocating for diverse leadership in enterprise architecture.
The Artificial (Un)Intelligence Conference is a global, 24 hour live online conference. Registration is free, and includes access to all sessions, including on-demand, at the conclusion of the event. No sales pitches, no marketing. Just the good stuff.
We're trying to find the unheralded people around the world who are doing cool things with AI. Is that you?
Check out our speaker gallery and then let us know what you're working on. We look forward to hearing your story.
More speakers added daily. Register to get the latest update.
Comments