The End of Software as We Know It
Arthur Mensch, the CEO of French AI lab Mistral, has a stark prediction for the tech industry. He believes AI is on the verge of replacing half of all existing enterprise software. This is not a gradual evolution. It is a fundamental shift in how companies acquire and use technology. Mensch suggests the era of buying static, one-size-fits-all software is ending.
Instead of purchasing subscriptions to massive platforms, companies will generate their own applications on demand. A manager could simply describe a business problem. An AI would then write and deploy a custom application to solve it instantly. This moves the value away from the pre-built software itself and toward the AI that can create it.
The current Software-as-a-Service (SaaS) model is built on selling seats and long-term contracts. It often leads to bloated products packed with features most users ignore. Mensch’s vision flips this entirely. It promises hyper-specific tools built for a single purpose, used for as long as needed, and then discarded. It’s a world of disposable, perfectly tailored software.
What This Means for Your Career
This potential future redraws career maps across the tech industry. Roles tied to the traditional SaaS sales cycle face the most immediate threat. If companies generate software instead of buying it, the need for sales reps who sell licenses and manage renewals diminishes significantly. The job shifts from selling a product to selling a capability.
Product managers will also need to adapt. The skill of crafting a five-year product roadmap becomes less relevant. The new premium will be on precisely defining business needs. A product manager's main job could become translating a complex operational challenge into a clear prompt for an AI. This requires a deep understanding of Business Process Reengineering and the ability to think in systems, not just features.
For software engineers, the work moves up the stack. Building the front-end of another project management tool is a less secure path. The real demand will be for those who can build and maintain the AI models that generate these apps. This puts a major focus on skills like AI/LLM Engineering & Fine-tuning. Another critical new role will be the AI integrator, a specialist who connects these generated apps with a company's core systems. This makes expertise in AI Workflow Integration incredibly valuable.
Finally, a new discipline of quality control will emerge. How do you trust an application written by an AI in seconds? Companies will need experts who can validate the logic, security, and reliability of generated code. This creates a demand for professionals skilled in AI Output Verification, a role that blends QA with an understanding of model behavior.
What To Watch
This shift will not happen overnight. The first signs will appear inside large tech companies. They will build internal platforms that allow their teams to generate simple tools for data analysis and workflow automation. Watch for announcements about these internal “app factories.” They will be the testing ground for this new model.
The next step will be the emergence of startups that offer this as a service. Instead of selling a CRM, they will sell access to an AI that can build any CRM you can describe. Also, pay attention to how existing low-code platforms begin to incorporate true generative capabilities. They are positioned perfectly to lead this transition. Hurdles like security, data privacy, and regulatory compliance are still significant. But the underlying technology is advancing faster than most people realize.