AI is radically changing custom software development, but not in the way people often think.
The key point is not that developers are being replaced en masse, but that the economic logic of software companies is changing. McKinsey has already concluded that software developers can perform certain tasks up to twice as fast using generative AI. These gains are primarily seen in documentation, code generation, and refactoring; for highly complex tasks, the effect is clearly more limited.
That distinction is crucial for the M&A market. As a result, value in custom software companies is shifting away from pure programming capacity and toward architecture, domain knowledge, quality assurance, security, and the ability to effectively integrate AI into the software development life cycle. AI can accelerate many processes, but at the same time it increases the importance of human oversight, clear review processes, and consistent engineering standards. McKinsey also emphasizes that AI can only marginally improve code quality and is most valuable as a complement to developers, not as a replacement for them.
This creates a new distinction for buyers. There are software companies that simply use copilots to deliver faster, and there are companies that have already redesigned their entire development process: better onboarding, shorter cycle times, greater reuse of components, faster prototyping, and a stronger link between product, engineering, and data. It is precisely this second category that is becoming more attractive in M&A. This aligns with the broader McKinsey finding that software engineering is among the functions where organizations most frequently report concrete cost benefits from AI.
At the same time, technical due diligence is becoming more complex. As AI code accounts for a larger portion of the output, issues related to IP, licensing, security, traceability, and maintainability are becoming more important. A buyer will want a clearer understanding of how code is generated, tested, validated, and documented. Not because AI reduces the value of software companies, but precisely because the gap between professionally integrated AI processes and ad-hoc use is rapidly widening. This conclusion follows from McKinsey’s observation that AI primarily increases productivity for specific types of tasks and that quality assurance remains dependent on human expertise.
We therefore expect that AI will not make the custom software market uniformly cheaper, but rather more selective. Companies that use AI solely as an efficiency tool will likely feel price pressure. Conversely, companies that leverage AI to achieve a demonstrably better delivery model, higher margins, and greater scalability will build a stronger equity story. For investors and strategic acquirers, the question will then shift from “How many developers does this company have?” to “How mature is this company in AI-enabled software delivery?”
