Is AI about to fracture the software market?

The software market is fracturing into two distinct worlds, and it’s being accelerated by Artificial intelligence. It is having more impact on software than any previous technology shift, and it’s happening before our eyes. Who will survive this shift, and what does it mean?

One side of the software world consists of edge applications. This includes workflow tools and SaaS layers that abstract data without owning it. The other side comprises rigid, mission-critical systems of record. These systems actually run financial services, government, and national infrastructure.

Enterprise Times talked with Richard Davies, UK Country Manager at Netcompany, and a veteran of the industry, about what this means. Davies argues that while AI excels at optimising the former, it poses an existential risk to organisations that attempt to embed it into the latter without a fundamental transformation. The market is splitting because the technology behaves differently depending on where it lives.

The Edge of the Organisation

A proliferation of software has emerged at the edge of organisations over the last fifteen years. These tools abstract data, manage workflows and sit on top of core systems. These systems do not own the data. They do not own the policies or the business logic. They simply abstract the information.

While they have given users access to data that previously would have been limited, the apps are not owned by the organisation. Instead, they are rented, and organisations have little say over new features and capabilities.

This architecture makes them susceptible to disruption from AI. Davies said, “If you’re just dealing with workflow and tools, then AI, as we currently understand it, is very adept at improving that by some order of magnitude.”

That is because the technology excels at these repetitive, rule-based tasks. It can speed up code development, automate testing and analyse vast datasets that humans cannot process.

This capability creates a false sense of security. Companies see the efficiency gains at the edge and assume the same logic applies to their core systems. This assumption is dangerous.

The Reality of Systems of Record

Core systems run organisations. They exist in financial services, in government, and power large national critical infrastructure. These systems are not based on SaaS products and do not run on abstracted layers. They hold the business logic and manage ownership and handoffs.

This is not a new problem. In the 90s, there was a push for better access to data and faster software development. Business units felt they were losing opportunities because they couldn’t access and analyse data. When the dot-com boom and the first versions of SaaS arrived, the conversation turned to integration and abstracting the data layer.

Data fabrics have become the way that organisations have attempted to separate data from applications. In 2010, we saw the arrival of Big Data, which required a change of architecture. By 2015, Machine Learning was the next big thing. For core systems, little has changed. Data is still dirty. Users don’t trust that it is ready for AI, and data maturity is still a work in progress.

But all those changes assumed that SaaS was the user interface and focused on that. For many organisations, it has become the key to business transformation. But does that make sense? Not according to Davies. He said, “You cannot SaaS your way to business transformation. It is just lunacy. Because you’re selling the Kool-Aid.”

The Wedge Driving the Market Apart

AI is driving a wedge between these two scenarios. Edge systems are ripe for disruption as organisations look to automate more. Additionally, the arrival of agentic AI, where AI does more than just automate simple tasks, is creating bigger changes.

Core systems require a different approach to get the most from AI. It must be embedded as part of those core systems that have never been designed for that.

Davies explains that AI in the marketplace is good for linear tasks. Speeding up code development and testing is one thing. But putting it into core systems requires deep industry understanding. You need to know how everything hangs together. You need auditability, and you need governance.

He continued, “It is not possible as we currently sit here. You need to transform those systems to be able to take advantage of some of the things that are being done at the edge.”

Organisations face a massive risk if they try to embed AI into current systems of record without transformation. The systems lack both the architecture to handle the complexity and the visibility required for safe automation.

A Reckoning for Consultancies

The rise of AI exposes a vulnerability in the business model of major consulting firms. For years, these firms generated revenue by configuring SaaS packages. They managed repetitive, rules-based tasks, and importantly, they sold the ease of configuration. AI reduces the value of rules-based repetitive work because it does this faster and cheaper.

According to Davies, “If most of your experience and your people are locked up in configuring systems that are heavily rules-based, then that would be a concern to me.”

Consultancies must evolve. Those who focus on design, testing, and understanding how to build platforms will survive and thrive. They understand that transformation requires building the platforms that support the business. They also know that transformation is a continuous process, not a one-and-done. That means you have to be able to audit and have deep observability of how systems and processes work

The market splits into three groups of consultancies. Pure business consultancies understand business models and processes. Technical consultancies understand system levels but have drifted into configuration. Hybrid models exist, but few are properly integrated.

Davies commented, “It’s problematic if you don’t have the ability to build systems. If your background is not building deep industry knowledge and building systems on the back of that, you will be disadvantaged in the world that we’re going into.”

The winners will be those who can merge business logic with technical architecture. They must build systems that have deep domain knowledge embedded in them. Configuring a SaaS product is not the same as reconfiguring a business process on a new platform.

The Role of Digital Twins and Modelling

Transformation requires visibility because you cannot change what you do not understand. This is where Davies sees the role of digital twins in this process. These models allow companies to simulate changes without risking production environments.

He said, “You have to have that understanding. Otherwise, you’re going to end up with a two-speed world where the technology and the cycles of AI will continue, but you’ve got nothing really to run it on that will make the game-changing impacts.”

This modelling capability is critical for safety. For vertical markets such as transportation and finance, mistakes can have catastrophic consequences. The ability to predict how a change in one part of the system affects the whole is the only way to manage complexity.

AI can now do this discovery work. It can map the data model. It can understand the application model. But it needs to be guided. It cannot just be let loose to re-engineer systems without oversight.

The Shift to Business-Led Process

The process is no longer set by IT or the core systems. The process is set by the business unit, and they are doing so at speed. Those who have learned to use AI or low-code tools to solve problems are the drivers. And much of the process they are creating is done through the code they produce. That creates a challenge.

Organisations must work out how to extract the business process from the code. It means a shift from the mentality where the process is created in modelling tools and is then used to define code behaviour. It is now dynamic. Code now drives process in many departments.

Davies agrees. “You have to build the systems around the business logic. And that’s in many ways, to your point, not always been the case.”

Business teams are creating their own solutions using tools like Claude to write code for their departments. This logic spreads through the business, and companies must find a way to capture this logic. What they lack, and what they need, are the tools to extract the process from the code.

Transformation is not an IT problem, nor is it just a business problem. It spans both. Those who will be successful in transforming core systems must use both.

The need for Digital Sovereignty

No conversation about the future of technology can avoid the issues of geopolitics and data sovereignty. The US Cloud Act and the influence of American tech giants create risks for European and UK organisations. This is pushing Governments to prioritise platforms that offer true ownership and control.

Davies noted, “You would be remiss in the geopolitical world that we live in, if you weren’t at least cognisant of where your data sat and who owned it.”

Mission-critical infrastructure has relied on third-party systems where vendors hold the keys. The changing nature of geopolitics has broken that trust. The UK government talks about digital sovereignty, but still buys American systems. This creates a vulnerability. If a US vendor is forced to shut down or freeze data, the impact could be severe.

Davies notes that the German state of Schleswig-Holstein is eliminating some US vendors. This trend will likely grow. Boards are starting to recognise the risk. Three years ago, they accepted third-party configurations. Now, they demand ownership and control.

The Path Forward for CIOs and Boards

CIOs and COOs face a difficult choice. They must decide whether to continue with edge-level optimisation or undertake the transformation of core systems. The latter requires a partnership between business and technology. It demands a deep understanding of domain logic. It requires a commitment to building platforms that can evolve safely.

Davies says, “Digital transformation is hard, and it’s iterative. You don’t just build it and walk away.”

The split in the software market is inevitable. Those who don’t transform core systems for AI will fall behind those that do. It will create a two-speed world where innovation happens at the edge, but the core remains stagnant.

Those who are successful will build platforms with embedded business logic. These platforms will be repeatable. They will have audit visibility and governance. Critically, they will also be able to accommodate the next wave of technology.

A Call for Realistic Expectations

The hype around AI is too often focused on how it will replace workers. But that ignores the knowledge that those workers have, and where AI and humans come together.

Davies commented, “It’s taking out jobs in highly repeatable product companies, but it’s not necessarily taking out jobs yet when you’re trying to transform and build net new platforms for specific mission-critical industry, reliant systems.”

The market noise is high with many vendors claiming to have the solution. They sell a promise of instant transformation. The reality, however, is much slower. It requires deep understanding. It requires a willingness to invest in the core.

Enterprise Times: What does this mean

The software market is dividing. The edge will be automated, and the core will be transformed. The gap between the two will widen. If organisations don’t want to be obsolete, they must act now.

They must build systems that can handle AI safely and embed domain knowledge into their platforms. They must ensure digital sovereignty. Most importantly, they must involve the business in the transformation.

The question for every leader is no longer if they will transform. The question is whether they have the depth of understanding required to survive the split.

As Davies says, “If you don’t have that basic set of tools and skills and capabilities and understanding, then you’re going to end up trivialising it all.”

The time for configuration is over. The time for transformation is here.

The post Is AI about to fracture the software market? appeared first on Enterprise Times.