AI Won’t Fix Poor Engineering. It Will Reward Good Engineering. 

Few technologies have generated as much excitement, investment and expectation as generative AI. In less than three years, it has evolved from an emerging capability into a board-level priority. Every major software vendor now has an AI strategy, engineering teams are experimenting with AI-assisted development, and organisations across every sector are searching for ways to improve productivity, reduce costs and accelerate innovation. 

Much of the discussion has centred on one question. How much faster can software be developed? 

It is an understandable question, but perhaps not the most important one. 

Generating code has never been the sole constraint on successful software delivery. Designing resilient architectures, understanding business processes, integrating complex systems, assuring quality, managing risk and deploying software safely into production have always consumed a significant proportion of delivery effort. Those activities do not disappear simply because AI can write code more quickly. 

The organisations that realise the greatest value from AI are unlikely to be those producing the largest volumes of code. They will be those that improve the effectiveness of the entire software delivery lifecycle.

AI changes the economics of software delivery 

For decades, software engineering has faced an uncomfortable reality. While development tools have continually improved, the complexity of enterprise technology has grown even faster. Organisations have added new applications, integrated acquisitions, expanded data platforms and responded to changing regulation, often without simplifying what already exists. 

The result is that many engineering teams spend a surprising amount of their time navigating complexity rather than creating new value. Technical debt accumulates, testing becomes increasingly time consuming and delivery pipelines become slower as organisations attempt to reduce operational risk. 

Generative AI has the potential to alter that equation. 

Developers can produce initial code more quickly. Documentation can be generated automatically. Test cases can be created in minutes rather than days. Legacy code can be analysed at a scale that would previously have required weeks of manual effort. Knowledge that previously resided with a handful of experienced engineers can begin to be captured and shared more effectively across delivery teams. 

These are meaningful advances, but they represent only part of the opportunity. 

The real value of AI lies in reducing friction across the entire software development lifecycle, allowing organisations to shorten the time between identifying a business need and delivering a high-quality solution into production.

More code does not automatically create more value 

One of the assumptions that has emerged alongside AI is that organisations generating the greatest volume of code will create the greatest competitive advantage. 

The opposite may prove true. 

Enterprise technology has never struggled because organisations produce too little software. Many large organisations already manage millions, and in some cases billions, of lines of code supporting critical operations. The challenge is maintaining, understanding and evolving those systems safely while continuing to deliver new capabilities. 

If AI simply enables organisations to generate more code without improving architecture, governance or engineering quality, there is a risk that technical debt will accumulate even faster than it does today. 

The objective should therefore not be writing more software. It should be creating better software that is easier to understand, easier to maintain and easier to evolve. 

That requires organisations to think about AI differently. 

Rather than viewing it as a coding tool, they should consider how AI can improve requirements analysis, solution design, testing, deployment, documentation, knowledge management and operational support. The greatest productivity gains are likely to come from reducing the friction between these activities rather than accelerating only one stage of the lifecycle.

AI exposes the strengths and weaknesses of engineering teams 

The organisations making the fastest progress with AI generally have something in common. Their engineering foundations were already strong before AI arrived. 

They have well-structured architectures, disciplined engineering practices, mature testing frameworks and software delivery pipelines that are designed for continuous improvement. AI allows these organisations to move even faster because it amplifies capabilities they have spent years developing. 

Conversely, organisations with fragmented technology estates often experience a different outcome. 

AI-generated code still needs to integrate with legacy systems. Automated tests still depend on stable environments. Intelligent agents still require reliable data. Modern development practices still rely on clear ownership, effective governance and engineering discipline. 

Where those foundations are weak, AI frequently reveals existing weaknesses rather than solving them. 

This is one of the reasons some organisations are finding that AI pilots demonstrate impressive results while enterprise-wide adoption proves considerably more difficult. Scaling AI successfully depends less on the sophistication of the model and more on the maturity of the engineering environment into which it is introduced.

Engineering excellence is becoming the competitive advantage 

This shift is particularly relevant in sectors such as financial markets and energy trading, where technology directly influences commercial performance. 

Trading platforms must evolve continuously in response to changing markets, regulatory requirements and customer expectations, all while maintaining exceptional levels of reliability. AI offers considerable opportunities to accelerate engineering activities, but those opportunities can only be realised where software delivery is already operating effectively. 

At Digiterre, we increasingly see clients moving beyond conversations about AI tools towards broader discussions about engineering capability. They recognise that AI is only one part of a much larger transformation. To realise sustainable value, organisations must also modernise legacy platforms, reduce technical debt, improve software quality and create delivery models that support continuous change. 

This is where AI becomes transformational. It is not because it replaces engineers or eliminates complexity. It is because it allows experienced engineering teams to spend less time on repetitive activities and more time solving the difficult problems that genuinely create business value. 

The organisations that will benefit most from AI

Every significant technological shift creates winners and losers. 

Previous generations invested in enterprise resource planning, cloud computing and digital transformation. Today the focus has shifted to AI. While the technologies themselves differ, the underlying lesson remains remarkably consistent. 

Technology creates competitive advantage only when organisations have developed the capability to use it well. 

The organisations likely to benefit most from AI over the next decade will not necessarily be those making the biggest investments or announcing the boldest strategies. They will be those that combine AI with disciplined engineering, modern architectures and delivery practices designed to adapt continuously as technology evolves. 

At Digiterre, we believe AI should be viewed as an opportunity to rethink the entire software delivery lifecycle rather than simply accelerating coding activities. Organisations that take this broader perspective are already seeing improvements in delivery speed, software quality, operational resilience and time to market. 

If your organisation is exploring how AI can improve software engineering, the most valuable question may not be which model to deploy or which coding assistant to adopt. The more important question is whether your engineering foundations are ready to extract the full value that AI can offer. In our experience, that is where the greatest long-term competitive advantage is created.