The European regulation on artificial intelligence (AI Act), which came into effect on August 1, 2024, redefines the strategic priorities of companies well beyond the mere adoption of tools. We are witnessing a clear shift: regulatory compliance becomes a guiding axis on par with growth or profitability. For leaders monitoring business developments in 2024, this framework alters operational trade-offs in the short and medium term.
AI Act and company obligations: what AI compliance changes
The AI Act does not only concern providers of artificial intelligence solutions. Any company deploying a tool that incorporates AI is potentially affected, including when using software purchased from a third-party vendor. A customer service chatbot, a scoring module in a CRM, a fraud detection tool: each of these use cases requires an inventory and qualification of the risk level.
Some prohibitions and the obligation to train individuals using AI apply from February 2, 2025. In practical terms, this imposes three simultaneous tasks:
- Map all tools and modules integrating AI into business processes, even those purchased as turnkey solutions.
- Qualify each use according to the risk categories defined by the regulation (minimal, limited, high, prohibited).
- Document human oversight, the traceability of automated decisions, and the quality of the data used.
SMEs that view AI as a mere marketing lever without addressing these obligations expose themselves to sanctions. We recommend integrating this aspect into the budget forecast, not at the end of the fiscal year.
Analyses published on the business site of Magazine Finance regularly detail the financial implications of these new regulatory constraints for SME and mid-sized company leaders.
Recruitment and high-risk AI: a blind spot in business in 2024
HR uses of artificial intelligence represent a specific risk angle that most articles on business trends overlook. Automated sorting of applications, performance evaluation, and certain pre-selection tools fall under the category of high-risk uses as defined by the AI Act.
This changes the game for any company using a recruitment platform that incorporates algorithmic scoring. Technical documentation, mandatory human oversight, and decision traceability are no longer optional. A candidate-job matching tool that filters profiles without documented human intervention exposes the company to direct legal risk.
In practice, we find that many HR departments adopted these tools during the accelerated digitalization phase post-2020 without ever auditing their operation. Auditing algorithmic HR tools becomes an operational prerequisite, not a project to postpone.
Company failures and cash management: the weak signal to watch
The dynamism of business creation in France coexists with a marked increase in failures in 2024. This combination reflects a structural phenomenon: launching an activity remains accessible, but survival beyond the initial years becomes complicated.
Cash flow is the primary discriminating factor. Companies that fail often share a similar profile: decent revenue growth, but insufficient margin monitoring and poorly anticipated working capital. The classic trap remains the gap between customer receipts and supplier payments, exacerbated by the lengthening payment terms observed in several sectors.
For entrepreneurs in 2024, three levers make the difference:
- Implement a weekly cash flow dashboard from the first month of activity, not at the first incident.
- Negotiate supplier payment terms before signing the first contracts, when the balance of power is still even.
- Plan a cash reserve covering several months of fixed costs, taking into account the actual collection timelines.
Preventing financial difficulties is not spectacular, but it is the parameter that separates enduring companies from those that disappear quietly.
Job skills and AI training: a non-negotiable investment
The obligation to train on AI tools, outlined in the AI Act timeline, is not limited to generic awareness. Each employee using an AI system must understand its limitations and potential biases. This requirement also applies to marketing teams configuring automated campaigns and salespeople relying on customer scoring tools.
Training is not about learning to “prompt” a language model. It focuses on the ability to assess the reliability of an AI-generated result, to identify cases where human intervention is required, and to document the decisions made. For companies selling online services or managing platforms, this skill becomes a measurable competitive advantage.
We observe that organizations that have anticipated this training aspect integrate regulatory constraints better without slowing down their activity. Training before deployment costs less than correcting after a formal notice.
Business in 2024 hinges as much on mastering regulatory frameworks as on product innovation. Companies that treat AI compliance, financial rigor, and skills development as topics for top management gain an edge over those that delegate them to support functions.
