By: Michael Privat
On weekends, coding feels like a different job entirely. There is no ticket queue waiting in the background, no approval chain to move through, and no quiet pressure to align with five other systems before anything can ship. It is just a problem, a clear starting point, and the freedom to see it through. In a few hours, something that did not exist is up and running. It works. It can be tested, adjusted, and improved in real time.
Then Monday comes around, and the same kind of work enters a structured environment. That contrast is most obvious during the week, where I work as Chief Data and Engineering Officer at Availity, supporting systems that connect health plans, providers, and patients across the U.S. at national scale. The engineers are just as capable. The tools are often identical. The goals are not smaller. Still, progress stretches in a way that feels disproportionate to the task itself. What should take a few hours quietly expands into days. The difference is not talent; it is the system surrounding the work.
AI has made it easier than ever to build. What it has not done is simplify the environments where that building happens. In many organizations, the effort required to create something has dropped, but the number of systems involved has not followed the same path. If anything, it has grown, and that gap is starting to show.
When Progress Starts to Look Like Accumulation
Most AI strategies do not begin with bad decisions. A new tool is introduced to solve a clear problem. Another is added to connect it with existing systems. Additional layers come in for monitoring, security, and optimization. Each step feels responsible, and each one has a clear rationale behind it. No one sets out to build something unnecessarily complex.
Over time, though, the system becomes heavier than expected. Every platform introduces its own dependencies. Every dependency creates another point of coordination. Instead of focusing on improving what is being built, teams spend more time making sure everything continues to work together. The architecture slowly shifts from supporting progress to demanding attention.
This is no longer just an internal frustration. It is showing up at a broader level.
Across Asia Pacific, organizations that expanded their AI environments aggressively are now moving toward consolidation because the operational complexity has become harder to manage than the problems AI was meant to solve (Source: IDC, 2025). At the same time, the conversation around AI has shifted away from what models can do in theory and toward how systems actually perform in practice, especially when it comes to integration, security, and measurable outcomes (Source: Glean, 2025).
There is also a growing urgency around speed. Nearly four in ten businesses now identify faster time-to-market as the main reason they are investing in AI (Statista, 2024). That makes every additional system more than a technical choice. It becomes a delay between idea and execution.
This is where a more difficult question starts to surface. What if progress is not about adding more, but about removing what no longer needs to be there? AI has reached a point where it can replace entire layers of software that were once necessary. Workflows that used to require coordination across multiple teams can now be handled in a single, continuous flow. Systems that existed mainly to connect other systems can often be simplified or eliminated altogether.
Even with that shift, many organizations still treat AI as something to stack on top of what already exists. New capabilities are added without rethinking the foundation. Legacy systems stay in place, and new layers are introduced to manage the complexity that follows. On the surface, it looks like progress. In practice, it often slows things down.
Data moves through more steps than it needs to. Visibility becomes fragmented. Small issues take longer to resolve because it is harder to trace where they began. The contrast becomes clear when looking at teams that move quickly. Early-stage companies often operate with fewer systems and fewer constraints. Their advantage does not come from better tools. It comes from having less standing in their way.
When Complexity Outgrows Clarity
There is always a stage where adding one more layer feels manageable. The system still makes sense. The team can explain how everything fits together, and work continues to move, even if it takes a bit more coordination than before. Nothing feels broken yet. Then, gradually, that clarity starts to slip.
The system grows beyond what any one person can fully understand. When that happens, progress begins to change. Tasks take longer, not because they are more difficult, but because the path from cause to effect is harder to follow. Issues appear in places no one expects. Confidence drops because the system starts to feel unpredictable.
AI accelerates this shift in a way that is easy to underestimate. Teams can now build faster than they can fully reason through what they are building. That gap creates risk, and it builds quietly over time.
This is why more organizations are starting to prioritize consolidation. A system that remains understandable is easier to maintain, easier to secure, and easier to evolve. When ownership is clear and data flows are direct, teams can move with a level of confidence that is difficult to achieve in a more fragmented environment.
At that point, a more important question starts to surface. Every system inside an organization needs to justify its place based on the value it provides today. If a system exists because of a limitation that no longer applies, it becomes worth reconsidering. AI creates an opportunity to revisit decisions that were made under very different constraints, even if those decisions once made perfect sense.
This is not purely a technical exercise. It involves changing habits and rethinking workflows that teams have relied on for years. There is a natural resistance to removing something familiar, especially when it has been part of how work gets done for a long time. Still, the payoff is difficult to ignore.
When unnecessary systems are removed, communication becomes clearer. Execution becomes faster. There are fewer points where things can fail, which makes the overall system more resilient.
The impact goes beyond operations. It changes how work feels on a daily basis. Teams spend less time navigating the system and more time actually building within it, which is where progress was always supposed to happen.
What a Strong AI Roadmap Actually Requires
A strong AI roadmap does not end in a more complex system. It leads to a simpler one. That usually shows up in very practical ways. There are fewer platforms involved in getting work done. There are fewer dependencies that need to be managed. There are fewer handoffs between teams before something moves forward. This kind of simplicity does not reduce capability. It reflects a more deliberate use of AI to remove what is no longer necessary.
Organizations that move in this direction start to operate differently. Teams spend less time coordinating and more time delivering. Decisions move faster because there are fewer barriers between an idea and its execution. Quality becomes easier to maintain because the system is easier to understand and control. Reaching that point requires a shift in how progress is measured.
Instead of asking what else can be added, it becomes more useful to ask where complexity is slowing things down. One of the simplest ways to see this clearly is to map how an idea moves from concept to delivery. When that path is laid out, the number of platforms involved, the number of handoffs required, and the places where delays tend to happen become difficult to ignore.
From there, the work becomes straightforward, even if it is not always easy. Teams can start identifying where AI can replace manual steps, where integrations can be simplified, and where entire systems can be removed without losing value. Some of these changes are technical. Others involve rethinking habits that have been in place for years.
The goal is not to build more. It is to carry less. When organizations focus on reducing complexity, they create an environment where AI can actually deliver on its promise. The result is a system that moves faster, adapts more easily, and supports the kind of execution that modern technology was supposed to make possible in the first place.
Michael Privat is a Chief Data and Engineering Officer leading a global team of 700+ engineers. With 25 years in tech, he helps organizations improve speed, clarity, and accountability. He helps stalled engineering teams become high-performing systems built on ownership, discipline, and modern AI-driven workflows through his “accountable autonomy” model. Follow him on LinkedIn or check out his Substack for more.
