Seerene News/Research

Business Success with Large-Scale Software Development

Written by Brandon M. Lewis | Jul 24, 2025 3:33:23 PM

In today’s enterprise, software development is no longer the exclusive domain of tech companies. From logistics to manufacturing to financial services, every industry has become a software industry. Yet despite billions invested into agile transformation and digital capabilities, most large organizations are still flying blind when it comes to the actual efficiency of their software production.

That’s dangerous—especially now that generative AI has entered the scene.

Software Is Eating the World—and AI Is Eating Software

Back in 2011, Marc Andreessen famously said “Software is eating the world.” He was right: most innovation today—whether in customer experiences, operations, or products—is delivered via software.

In 2017, Nvidia CEO Jensen Huang added a twist: “AI is eating software.” At the time, he meant AI was transforming areas like computer vision and medical diagnostics. Today, it’s transforming how software itself is built.

With the advent of LLMs like GPT, Claude, and Gemini, generative AI has moved into the software development process itself. Developers now have copilots. Code is being written, optimized, and even refactored by machines.

This raises a critical question for CIOs: Will AI replace developers?
No. But it will replace bad ones. And it will expose inefficient software organizations.

Two Worlds of Software

There are two kinds of software systems:

  1. Project-Based Systems: These are classic waterfall systems with a clear beginning and end (e.g., SAP deployments). For these, generative AI will increasingly automate much of the code implementation—threatening traditional system integrator models and shifting the business toward configuration over customization.
  2. Evolving Systems: These include proprietary applications, embedded systems, and unique digital products. They are never truly finished. Specifications change constantly, often driven by evolving user behavior or strategic pivots. These systems require continuous adaptation—an ongoing architecture of change.

AI can assist in both. But only in the second type will the human developer—and the organizational structure around them—remain essential. The catch? These evolving systems are also the ones where most organizations have no idea how efficiently they are operating.

Generative AI Will Expose Your Waste

Seerene has measured the impact of GenAI tools across major organizations. Here’s what we’re seeing:

  • Code throughput triples. Developers generate significantly more code when supported by GenAI.
  • Defect fixing spikes. Without robust test automation, this productivity surge causes a tsunami of bugs, which eats up all the time saved.
  • Code complexity grows. Poor use of GenAI leads to deeply nested, hard-to-maintain code—creating long-term technical debt.

The result? A false sense of productivity. And worse: a growing gap between developer output and actual business value.

This is why CIOs can no longer afford to manage software development without hard data. Transparency is no longer optional—it’s a prerequisite for responsible leadership.

The Scarcity is Not Code. It’s Developer Time.

Your developers are your most valuable production resource. But in nearly every large enterprise Seerene has analyzed, more than 70% of developer time is lost—to refactoring, firefighting, legacy debt, or inefficiencies in the development process.

What’s shocking is not that this waste exists—but that most CIOs have no reliable way to measure it.

Now that GenAI is accelerating everything—good and bad—this blind spot becomes fatal.

Strategic Steering for the AI Era

Generative AI offers real value, but only when introduced with:

  • Upskilled developers who can use it responsibly.
  • Robust architecture to absorb AI-generated complexity.
  • System-wide measurement of software production KPIs.

That last one is key.

CIOs must move beyond toolchain visibility and toward executive-grade transparency: strategic metrics that show how efficiently each organizational unit is producing business value through software.

Done right, this creates not just oversight—but momentum. Transparency itself becomes a transformation engine. Teams start competing to improve. Efficiency becomes a cultural norm. And AI, rather than creating chaos, becomes a force multiplier for innovation.

The bottom line: AI won’t replace your developers. But it will replace any illusion you had that software production can remain unmanaged, unmeasured, or inefficient.

If you're a CIO preparing your organization for the AI era, ask yourself:
Do I know where my developer time is going?
And if not—how long can I afford not to?

Note: This article was inspired by a masterclass given by Dr. Johannes Bohnet at the "Transform or Face Disruption" executive exchange, the tenth event in a series by the Software Excellence Network. While I attempted to stay true to Dr. Bohnet's message, in order to understand the full depth of his remarks I suggest you watch the full presentation.