top of page

Artificial Intelligence and SMEs: Where Should You Start?

Writer: Paola Sabran
Paola Sabran
Oct 26
5 min read


“I keep hearing about AI everywhere, but in practical terms, what would actually be useful for my business?”


Artificial intelligence is now omnipresent in discussions aimed at businesses, where it is presented as a way to save time, improve productivity, make better use of data, and support decision-making.


Its use is also growing rapidly. According to the Baromètre France Num 2026, published by France Num, the French government initiative supporting the digital transformation of small businesses, 53% of SMEs report using at least one artificial intelligence solution, compared with 40% of all very small businesses and SMEs. It should be noted, however, that only 32% of SMEs use paid solutions.


Behind this rapid growth, however, lies a much more complex reality for small-business leaders. Solutions are multiplying, new features are appearing in existing software, start-ups are regularly launching new tools, and training options continue to expand.


Faced with this abundance of choice, it is becoming increasingly difficult to distinguish what is genuinely useful, sufficiently mature, and suited to the company’s needs. Business leaders can quickly find themselves without clear reference points, unsure which solutions to consider or where to begin.


This challenge is even greater because most companies have limited resources to devote to this process. Many have no IT department, no dedicated data team, and often no CFO to support them in making these choices.


In this context, the question is probably not which artificial intelligence tool to adopt first, but rather where AI could genuinely add value to the business.



Start with the Need, Not the Technology


Before looking for a solution, one question should come first: what problem are we trying to solve?


Within an SME’s Accounting and Finance function, needs can vary widely: reducing the time spent producing reports, gaining a better understanding of margins, quickly finding information across multiple documents, identifying anomalies, preparing summaries, improving forecasts, or automating some repetitive tasks.


The important point is to define the need independently of artificial intelligence.


If producing a monthly dashboard requires several hours of data collection, manual re-entry and reprocessing, the first issue to address is how information flows through the business and how it is used — a topic we have already discussed in relation to how well the company’s tools communicate with each other. Similarly, if the company does not know precisely which products, customers, or projects are the most profitable, it must first ensure that the necessary data is available and properly structured before attempting to analyze it with AI.


Artificial intelligence may then contribute to the solution, but it should not be used to bypass the underlying problem.


Likewise, AI can process large volumes of data very quickly, but it does not automatically turn incomplete or poorly structured information into reliable data.


The same principle applies more broadly to digital transformation projects: as discussed in previous articles, automating a poorly organized process or applying a powerful tool to unreliable data does not guarantee an improvement in the way the business operates.



First Look at What Your Existing Tools Can Already Do


The current enthusiasm surrounding artificial intelligence can create the impression that adopting AI necessarily means finding and adding a new tool.


That is not always the case.


Accounting, management, cash management, reporting, Business Intelligence, and financial planning software are progressively incorporating new automation capabilities and, increasingly, artificial intelligence features such as natural-language search, data analysis, anomaly detection, automated commentary, forecasting assistance, and support in interpreting results.


Before adding another application, it may therefore be useful to assess what the tools already in use can actually do and which new capabilities have become available.


This approach can help avoid unnecessarily multiplying software applications, subscriptions, and interfaces between different systems, while maintaining a more coherent working environment.


An additional solution may, of course, be justified when it addresses a need that is not already covered. Its contribution should then be clearly identifiable and significant enough to justify the cost, integration constraints, and the time required for implementation and adoption.


Some tools, known as AI agents, can also access several applications or data sources and automatically carry out sequences of tasks. For example, an AI agent could combine data from the accounting system, cash management software, and the CRM to identify customers with overdue payments, detect an upcoming cash flow pressure point, and prepare a summary of the issues requiring attention.


Whatever approach is chosen, confidentiality, personal data protection, trade secrets, and compliance with the French and European regulatory framework should be considered from the moment a solution is selected, rather than only after it has been implemented.



Test a Limited Use Case Before Expanding


For an SME, there is no need to begin with a complex project or a complete transformation of the organization.


A first use case can focus on a clearly identified need: summarizing documents, preparing commentary for a report, analyzing a set of data, finding information more quickly, or assisting with a repetitive task.


The purpose of the test should not simply be to establish that the tool works. It should also determine whether it genuinely improves the existing process.


Is the time saving significant? Is the result sufficiently reliable? How much checking is still required? Does the tool genuinely simplify the work, or does it simply shift part of the workload toward additional verification?


A task completed in a few seconds does not necessarily translate into a productivity gain if substantial time must then be spent checking, correcting, or rewriting the output.


The real measure should therefore be the overall improvement in the way the business operates, rather than the technical performance of the solution alone.



Do Not Confuse Technological Urgency with the Company’s Actual Needs


The speed at which artificial intelligence is evolving can make business leaders feel that they must act immediately or risk being left behind.


However, not adopting every new solution as soon as it appears does not mean falling behind.


The greater risk may instead be to multiply tools without a clearly identified need, connect sensitive data to solutions that have not been sufficiently assessed, or devote time and resources to technologies that ultimately do not improve the way the business operates.


Integrating artificial intelligence should therefore not be a technology project in itself. It should address a clearly identified need, deliver a measurable improvement, and fit within an organization that is sufficiently structured to manage both its uses and its risks.


For an SME, the objective is probably not to use artificial intelligence everywhere, but to gradually identify the situations in which it can genuinely create value, test solutions with appropriate judgment, and expand their use when the results justify doing so.


 

Key Takeaways

✓ Faced with the growing number of artificial intelligence solutions, the starting point should be the need or problem the company is trying to address, rather than the tool itself.

✓ Before adding another solution, it is worth assessing what the software already in use can do and whether the company’s data and processes are reliable enough to produce meaningful results.

✓ An AI use case should be assessed on the basis of its actual contribution to the business: time saved, quality of the results, level of checking required, costs, and implementation constraints.

✓ A gradual and targeted approach makes it possible to test the benefits while managing the confidentiality, security, and compliance issues associated with the company’s data.

 


Afine Conseil supports SME business leaders in organizing, digitalizing, and transforming their Accounting and Finance function, helping them identify the tools and developments best suited to their company’s needs.





Read Also


bottom of page