Doug Shannon is a global leader in digital transformation, specializing in AI, GenAI and intelligent automation.
Every generation has a habit of viewing new technology through the lens of the world it already understands. We see cars and imagine faster horses. We see computers and imagine faster calculators. Today, many organizations see AI and imagine better automation.
I think we are making the same mistake again.
As predictive models continue to evolve, we are moving beyond systems that simply respond to requests. We are entering an era where intelligence can continuously observe, reason, predict and recommend actions across the enterprise.
Much of the current discussion around AI focuses on productivity, co-pilots, agents, automation, orchestration and, yes, code generation. That is where most of the attention is today.
I think something bigger may be developing underneath it. The more interesting question is what happens when intelligence begins optimizing for the future instead of preserving the past.
For decades, organizations have accumulated technical debt, process debt, organizational debt and integration debt. What I am increasingly seeing is something different. At least, I do not think we have a name for it yet.
I call it interconnected debt. This is not a single application, platform, workflow or integration. It is the growing web of relationships, dependencies, assumptions, exceptions, policies and machine-generated decisions that connect the modern enterprise together.
Historically, complexity accumulated at human speed. Developers wrote code. Architects designed systems. Teams built integrations. Business processes evolved. Complexity grew slowly enough that humans could generally understand what they were creating.
That assumption is beginning to break.
AI-assisted development, autonomous engineering systems and increasingly capable reasoning models are changing the rate at which complexity can be created. Systems can now generate, modify, optimize and rebuild parts of themselves faster than humans can fully understand the downstream consequences.
Individually, none of those changes seem significant. Collectively, they create something new. That is interconnected debt.
In many ways, interconnected debt is the organizational equivalent of compound interest. Every new dependency, workflow, exception, integration or automated decision may appear rational in isolation. Over time, their interactions become increasingly difficult to model, predict and understand.
Unlike traditional technical debt, interconnected debt can accumulate at machine speed. The result is an environment where complexity grows faster than human understanding.
This is where predictive intelligence becomes interesting. Most people assume future enterprise AI systems will help maintain existing environments through better monitoring, stronger observability and fewer operational disruptions. I agree. What I think gets overlooked is what happens when intelligence develops a deep enough understanding of the environment itself.
Imagine a system that understands not only how the enterprise operates today but how it arrived there. Every integration, work-around, process exception, technology decision and organizational compromise becomes part of a larger model. At some point, that system may reach a different conclusion than the humans maintaining it.
For decades, enterprises have treated complexity as the unavoidable cost of growth. Every acquisition, modernization effort, compliance requirement and business decision added another layer to the environment. Over time, those layers became accepted as permanent.
Humans inherited complexity and learned how to manage it. Intelligence at scale may inherit complexity and question why it exists in the first place. Instead of asking how to maintain complexity, it may begin asking whether the complexity is necessary at all.
If the business outcome remains the same, why preserve every historical decision that led to it? That question is what makes the Gordian Knot relevant. In the original story, the knot was considered impossible to untangle. Alexander never attempted to unravel it. He simply cut through it.
Future predictive systems may eventually reach similar conclusions—not because they are conscious or acting independently but because they are optimizing against the objectives we provide. The most efficient path forward may not be maintaining complexity. It may be removing it.
There is another possibility that receives far less attention. If intelligence operating at scale begins rebuilding systems around new objectives, it may also begin creating new forms of interconnected debt. After all, this is not the first time complexity has emerged from optimization.
Many of the systems we maintain today were themselves attempts to solve the constraints of an earlier era. The Gordian Knot was not created overnight. It emerged one decision at a time until nobody fully understood the whole. Future intelligence may eventually cut through today’s knot only to create another one tomorrow. The difference is that the next knot may be forming at machine speed.
That is the part I think leaders should be paying attention to.
Complexity has always existed. The challenge has never been complexity itself. The challenge has been maintaining enough visibility to understand the consequences of change. Historically, those changes occurred slowly enough that humans could adapt. Architects understood the systems they designed. Developers understood the code they wrote. Organizations accumulated complexity one decision at a time.
What happens when those decisions occur thousands or millions of times faster? What happens when systems begin rebuilding systems? What happens when interconnected debt starts accumulating faster than human understanding can keep pace?
The risk is not that intelligence removes complexity. The risk is that it creates a new form of complexity that remains largely invisible until it becomes critical. After all, many of the systems we struggle to maintain today were originally built to simplify something.
Complexity is rarely designed—it accumulates. If intelligence operating at scale begins reshaping the enterprise, leaders may need to focus less on what is being removed and more on what is being created in its place.
The next Gordian Knot may not be the one we inherited. It may be the one we are building right now.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
