The digital transition cannot be managed with a checklist of tools to deploy. We observe that failing projects share a common trait: they treat transformation as an IT issue, whereas it involves governance, regulatory compliance, and data architecture. This article targets the concrete friction points that general guides tend to overlook.
Electronic invoicing and NIS2: two regulatory constraints that dictate the timeline
The French reform on electronic invoicing imposes an operational obligation on all companies subject to VAT, including those that have never formalized a transformation project. This regulatory timeline forces compliance: without a partner dematerialization platform, without structured flows between ERP and public portal, the company finds itself non-compliant.
At the same time, the European NIS2 directive (Directive (EU) 2022/2555, published on December 27, 2022) significantly broadens the scope of organizations subject to cybersecurity obligations. It strengthens risk management, incident notification, and supplier control. The European deadline for transposition was set for October 17, 2024.
We recommend treating these two topics as structuring milestones of the transition project, not as annexes. Mapping supplier dependencies, identifying sensitive data flows, and establishing an incident response plan: these actions condition the robustness of any digital component added later. Specialized resources support this approach, notably https://www.lepavenumerique.fr/, which addresses these digital structuring issues for companies in transition.
AI adoption in companies: a sectoral divide to integrate into the strategy
Artificial intelligence has become a distinct marker of digital maturity compared to other technologies. According to the 2024 ICT survey by INSEE, French companies with at least ten employees still showed limited use of AI, with a significantly higher adoption in the information and communication sectors.
This sectoral divide has a direct consequence on the transition strategy: applying an AI deployment plan designed for an IT services company to an industrial SME or retail business yields disappointing results. The level of data structuring, the internal culture of indicator-driven management, and the availability of technical skills vary significantly from one sector to another.
Prioritize use cases with measurable returns
Before investing in a generative or predictive AI tool, we recommend identifying a specific use case with quantifiable gains: automating support ticket sorting, forecasting restocking, document classification. A pilot project on a limited scope allows for validating the quality of available data and the teams’ readiness.
A failed AI pilot on poorly structured data costs more than a postponed project. It is better to invest first in cleaning and normalizing the repositories before connecting an algorithm.
Cybersecurity: a governance issue, not an IT budget line
NIS2 directly engages the responsibility of governing bodies. It is no longer a topic solely carried by the CIO before the executive committee. The directive imposes oversight obligations at the board of directors or management level, depending on the legal form.
Structuring the digital transition without integrating cybersecurity from the design stage is akin to building a building without an evacuation plan. We observe three recurring gaps in the projects we audit:
- Absence of mapping dependencies concerning cloud and SaaS providers, making it impossible to evaluate the supply chain risk required by NIS2.
- Incident response plans drafted after the deployment of tools, rather than being designed upfront to cover the new data flows created by the transformation.
- Cybersecurity budgets indexed to past revenue, without considering the increase in attack surface related to new digital components.
Cybersecurity must be treated as a governance issue with regular reporting to the management committee, risk coverage indicators, and an access review process aligned with each stage of digital deployment.
Change management and digital skills of teams
Tools do not transform anything if human processes remain unchanged. The support offer for digital transition has significantly expanded in recent years, but internal absorption capacity remains the bottleneck.
Change management is not limited to a training day on the new tool. It involves reviewing job descriptions, designating digital referents by department, and measuring actual adoption (usage rates, support tickets, persistent manual workarounds).
Build a tiered skills development plan
We recommend sequencing skills development into three tiers aligned with technical deployment:
- Basic tier: mastery of collaborative tools and dematerialized processes (invoicing, electronic signature, document management).
- Intermediate tier: reading dashboards, leveraging business data, participating in process reviews.
- Advanced tier: configuring automations, contributing to AI use cases, participating in internal security audits.
Each tier must be validated before moving on to the next. Deploying an AI tool to teams that do not yet master digital document management generates rejection and disproportionate support costs.
The digital transition of a company in 2024 hinges less on the choice of technologies than on three structuring axes: regulatory compliance as a timeline driver, cybersecurity as a governance discipline, and skills development as a condition for sustainability. Organizations that address these three topics in parallel with technical deployment are the ones that sustainably transform their processes.
