For a few months now, I’ve been noticing a trend that worries me a bit in development teams: generative AI has become so proficient at writing code that many believe the difference between good and bad developers will disappear. This is a mistake. Yes, ChatGPT, Claude , and other coding assistants are revolutionizing the way we develop. Yes, they accelerate repetitive tasks, testing, and corrections. But here lies the paradox: the easier AI makes code production, the more true value shifts elsewhere. It now resides in business understanding, intelligent prioritization of features, and the quality of customer support. This is what I’ve observed while working with dozens of software publishers, and this is what will truly make the difference in the next three years.
Summary
Producing faster does not mean producing better
I must be honest: in the software industry, we often confuse development speed with real value creation. This is understandable, as releasing a new feature is visible and measurable. Reducing a delivery time is quantifiable. But this confusion makes us forget something essential: software is not better because it contains more features. It is better when it truly meets the needs of users. 🚀
Generative AI, no matter how powerful, cannot arbitrate this fundamental question. A coding assistant can propose an elegant architecture, generate automated tests, or suggest a brilliant fix. But it does not know, on its own, if the feature requested by a client is truly a priority. The real risk, if we delegate too much to AI, is to produce faster, but mediocre things, poorly prioritized or too complex for real needs.
I have seen teams that have doubled their productivity thanks to AI tools, but who have also doubled the number of unused features in their products. This is the classic trap: automation without strategy creates noise, not value. The best publishers I know use AI to save time, but they reinvest that time in strategic thinking, not in frantic production.
True value comes from deep business understanding
Today, the value of a software publisher no longer rests solely on its technical ability to develop. It mainly relies on its capacity to understand what needs to be developed, why, for whom, and with what level of priority. For a B2B publisher, this requires a fine and nuanced understanding of real usage. Not the kind found in generic market studies, but the kind built through daily proximity with users. 💡
This is exactly what makes the difference between software that impresses during a demonstration and software that remains useful six months later, when teams are actually using it, with their urgencies, exceptions, and field constraints. Creating value is therefore not just about automating a task or adding a new feature. It is about understanding how your clients work on a daily basis, what decisions they need to make, and what data they manipulate.
This degree of business understanding cannot be invented and certainly cannot be generated by AI. It is acquired and refined over time, through direct exchanges, field feedback, and attentive observation. It is a sustainable competitive advantage, precisely because it cannot be automated. Publishers who invest in this understanding will create products that AI cannot easily replicate. We are already seeing this trend with the best analytics tools that combine data and business expertise.
Knowing how to prioritize becomes a major competitive advantage
All clients have requests. Many are legitimate, some are urgent, and others seem urgent but hide a deeper problem. In this situation, responding to the need does not always mean adding a new feature. The right answer may also be to leverage the existing differently, adjust a setting, revisit a user journey, or improve what is already there. This is a difficult exercise because it forces the publisher not to say yes to everything, right away. ⚖️
With AI, this exercise becomes even more critical. When producing code becomes easier and faster, the temptation is strong to develop and accumulate endless features. But in the end, the user will not judge software by the number of possibilities it contains. They will judge it by its ability to help them do their job more simply, quickly, or calmly. Intelligent prioritization in product development will therefore become a true marker of seriousness and maturity.
I already see publishers who understand this standing out. They say no to certain requests, they explain why, and they offer alternatives. Their clients respect them for that. It is the opposite of the “yes to everything” mentality that creates convoluted products, impossible to master and maintain. This strategic approach aligns with the principles of conversion rate optimization where every element must serve a clear objective.
Customer support: the true human difference
There is one point that many underestimate: customer support and assistance. The more powerful and sophisticated software becomes, the more it needs to be well explained, well configured, and well adopted to create value. AI can certainly help support by categorizing requests, suggesting answers, summarizing a ticket, and speeding up diagnostics. Let’s use it for that; it’s legitimate. 🤝
But this is just a tiny part of the customer relationship. Good support is not just about fixing a bug or answering a technical question. It is about understanding what is really blocking, identifying friction points, surfacing real needs, and supporting a change in usage. In a world where many things can be automated, the quality of human relationships becomes more visible and more valuable. Clients will quickly differentiate between a publisher who responds quickly with a standard AI-generated answer and a publisher who truly understands their problem and responds precisely.
This is also where you learn about the real unmet needs. This is where you discover use cases you hadn’t anticipated. This is where the business understanding I mentioned earlier is built. So yes, use AI to optimize support, but do not let it replace the human relationship. AI agents for automation must always be supervised by humans.
How the best publishers will use AI to amplify themselves
AI will not standardize the value of all publishers. It will primarily highlight which ones know where they are going and which ones are improvising. When used well, AI can help save time, process certain technical issues faster, and improve code quality. Good publishers will reinvest this time where it really counts: in listening to users, the quality of service, and the clarity of product choices. 🎯
I see a new category of publishers emerging who use AI as a productivity multiplier but maintain strategic control. They use AI tools to automate repetitive tasks, but they keep humans for important decisions. They use AI to speed up development, but they slow down prioritization to make it more thoughtful. It is a subtle balance, but it is the one that creates real value.
Conversely, those who let themselves be blinded by the promise of AI risk running faster in the wrong direction. They will produce a lot, but few useful things. They will have frustrated clients with overly complex products. They will waste time maintaining features that no one uses. This is a classic technology trap: confusing capability and strategy.
Conclusion
Generative AI is truly changing the way we develop software. It is a technological revolution, and I do not underestimate it. But this revolution will not create uniformity. It will create divergence. Publishers who understand that AI is a tool, not a strategy, will amplify themselves. Those who think AI will solve all their problems will be disappointed. The real difference is human judgment: knowing what to develop, why, for whom, and how to support it. This judgment cannot be automated, and it is this judgment that will make the difference in the years to come.
If you are a software publisher, the question is not “how to use AI to code faster.” The question is “how to use AI to better understand my clients and better prioritize my development.” This is a more difficult question, but it is the right question. And it is the one that will truly allow you to stand out.
In Brief
- AI accelerates development, but speed does not automatically create value
- The real difference lies in deep business understanding and intelligent prioritization
- Customer support and human relationships become more important, not less
- The best publishers will use AI as a multiplier, not as a replacement for strategic judgment
