Too many organizations confuse innovation with purchasing AI licenses. At Inside, we believe that success depends less on tools than on preparing teams to use them.
When digital disrupted our economies, the companies that prospered were not those that simply bought computers and software, but those that rethought their models. AI requires the same perspective. It is this observation that pushed us to structure our own approach via the creation of the iAxLab, a transversal unit playing both the role of locomotive and safeguard. The objective was to offer structured and secure access to AI tools, in order to unite energies and encourage innovation within a framework of trust, guaranteeing the protection of company data.
The trap of the tool in the face of the urgency of acculturation
Many companies are launching into AI through the exclusive prism of the tool. This is a fundamental error. Without governance, without a security framework and above all without culture, there is no sustainable industrialization possible. Acculturation is the keystone of the entire transformation. At Inside, we have succeeded in acculturating our workforce through concrete workshops. The trigger? Make each employee understand what AI will change, not in the world, but in their daily lives.
Yet adoption remains a challenge. If the enthusiasm is general, transforming the test is complex. For what ? Because actually using AI requires getting out of your comfort zone. It means agreeing to no longer open Google or Word as your first instinct, but to rethink your work process from scratch. AI must no longer be plan B when things get stuck, it must become the starting point for thinking.
Concrete use cases, far from magical speeches
Once the framework has been established (ethical charter, secure environments), theory must give way to pragmatism. It’s not about doing AI for the sake of doing AI, but about targeting tedious tasks where human added value is low. Internally, we have gone well beyond the simple gadget stage to anchor AI in the reality of our businesses. Our sales teams and consultants rely, for example, on a custom-developed CV generator (CV Gen), which automates the adaptation of profiles to the specific requirements of each client to save valuable time.
On the human resources side, we have designed a tool capable of combining the market context, the level of experience and our internal equity policy in order to offer the fairest remuneration. Finally, our pre-sales process has also been “augmented”: the use of solutions like NotebookLM now allows us to analyze imposing specifications in record time and to build solid commercial proposals, directly inspired by our past successes.
In IT development, we are already seeing time savings of around 15 to 20%. And the evolution does not stop there: with the arrival of AI agents, the very profession of developer will change. From “code producers”, they will gradually become “controllers and validators”. An essential letting go, but which requires, again, strong psychological and managerial support.
Sovereignty and ROI, the real challenge of tomorrow
We must also stop seeking at all costs an immediate and purely mathematical financial Return on Investment (ROI). It is necessary to integrate the notion of temporality: the ROI of AI, yes, but when? While it is undeniable that these tools save time in the short term, the question of their long-term profitability inevitably arises. Indeed, going twice as fast to generate unmaintainable applications or poor quality code is to accelerate your own failure by creating technical debt which will cost much more tomorrow. At the start, the ROI of AI is measured elsewhere: in the employee experience, in working comfort and in the quality of deliverables.
At the same time, we must not forget a major economic reality: AI is not free. Each request, each generation of content has a cost (the “comput”). Without strategic thinking upstream, a company can very quickly see its invoices explode and find itself at the opposite end of the ROI gain sought. This is why a FinOps approach applied to AI is now essential to monitor, optimize and control the consumption of these tools on a daily basis.
Finally, as cloud architectures mature, data sovereignty is no longer an option. This is the sine qua non condition for 2026. Our clients, particularly large accounts, demand total control of their data. Relying on European solutions (such as Mistral AI models) deployable in a closed environment or on internal infrastructures is becoming a major competitive advantage.
AI is an accelerator, you choose the direction
If I had to give just one piece of advice to a manager who wants to launch their company in AI today, it would be this: never underestimate the human factor. AI is just an accelerator that amplifies your existing reality. If your processes are bad and insecure, AI will only accelerate the production of flaws and poor outcomes on a larger scale. Secure your base and involve your IT department as a true partner to avoid Shadow IT: letting sensitive data and intellectual property escape to uncontrolled public tools is a critical risk.
Above all, train your teams, because if the historical professions are not going to disappear, they will profoundly change towards a supervisory position. The future belongs to “augmented collaborators”: The augmented Product Owner who can create concrete applications, immediately usable by his users. This approach makes it possible to conduct real tests, to precisely identify the business value generated, and thus to determine the relevant financial investments. Or the augmented recruiter focused on people and soft skills. Those who refuse to adapt, or who naively believe that purchasing a software license is enough to magically transform themselves, will remain on the side of the road.
