We have all heard about the accelerating pace of AI and the subsequent noise that accompanies it – both factual and misleading in equal measures. As well as AI’s impact on day-to-day life, AI is making waves in the software world too and this is accelerating rapidly, especially in coding and software development, with newer models significantly improving capabilities over the past year in coding, automation, and the ability to build quick prototypes.
This has caused substantial ructions in public markets with software/SaaS companies facing selloffs due to fears that AI could be a major disruption, resulting in significant corrections in recent business valuations.
However, enterprise software has enduring value because it contains much more than just code. Whilst AI might quickly be able to generate a specific app to challenge an existing market solution, any incumbent software business holds valuable assets, such as proprietary customer data, embedded workflows, domain/sector expertise, integrations, regulatory compliance and, importantly, customer understanding, trust and accountability. In short, the idea that anyone can easily replace complex enterprise systems with AI-generated tools is overstated and much harder than is suggested.
What is, potentially, more likely is that AI will significantly automate many white-collar workflows over the next 5-10 years, especially where these involve repetitive, administrative tasks. Rather than replacing existing software outright, it is more likely that existing software businesses can have a significant advantage if they integrate AI into their products, using their existing customer relationships, data, and domain expertise to build “agentic” features on top of their platforms.
Customers are typically risk-averse and generally do not want to replace critical systems with ad-hoc AI-built tools. They are more likely to prefer that existing software providers integrate AI into products that they, the customer, already trust.
In the software world this likely means that businesses will need to sacrifice profitability over investment. They need to invest now in AI product development, experimentation, and organisational change, even if that is to the detriment profitabilitytemporarily. This potentially involves prototyping AI features with customers, accepting that not every initiative will work, but knowing that speed of learning matters.
Those who don’t adapt risk losing value and market position, as there is no doubt that AI can help lower barriers for new entrants. That said, those that move quickly can expand beyond software into automating customer workflows, particularly those that are repetitive, administrative tasks, using their subject-matter expertise to their advantage.
The greatest risk for software businesses lies with those whose software products are broad, generic and low differentiation, without deep workflow knowledge or proprietary data. And if these businesses are slow to adopt AI, then they are at significant risk.
As in many spheres of business, the key usually lies in the customer relationship and the associated sector knowledge. So, whilst AI is genuinely transformative and likely to automate major parts of business operations in many areas, complex enterprise software is unlikely to be easily displaced because trust, compliance, governance, and operational reliability matter equally as much as raw technical capability.