AI in SMEs: an unattainable luxury?

Nine out of ten companies say they will increase their AI Readiness budget (data, platform modernisation and training) by at least 10%, while a third anticipate increases of up to 49% in items related to strategy and cultural change. However, 76% of organisations concentrate AI in only one to three use cases which are still in the pilot phase. These statistics, included in the study ‘Playbook for Crafting AI Strategy’ (MIT Technology Review), are certainly based on a sample dominated by large corporations with revenues exceeding USD 500 million. However, they present an underlying message that is relevant for SMEs in our territory: the barrier is no longer technology, but rather the ability to turn an idea into a realistic project, evaluate its return on investment, and scale up while ensuring legal and data security. In the study, small and medium-sized enterprises, who are more exposed to budgetary constraints, cite financing as their main obstacle to progress, well ahead of technical challenges. 

Another recent report (‘AI in the Enterprise’ by OpenAI) shows that AI generates value when it is used for critical processes, such as automated customer service, product search and tagging, and personalised communications. 

In the context of SMEs, the key lies in conducting rapid assessments based on limited data samples and focusing on clear, tangible metrics of success. Rather than creating complex, standalone solutions, SMEs should integrate AI capabilities directly into their existing platforms (e.g. CRM or e-commerce platforms), making full use of the flexible and affordable options available in on-demand consumption models. Similarly, quickly training key employees to act as internal experts in the adoption and use of these technologies accelerates results and minimises dependencies. 

Empirical evidence clearly shows that data quality and liquidity (the ability to combine and analyse data without friction) are the main factors limiting the speed of AI deployment. Therefore, it is imperative to map out friction, calculate the return on investment for 3–6-month periods, assess data maturity and design a minimum viable product (MVP). Bear in mind that, as with any digital transformation, the key to success lies in changing mindsets and culture, not just the technology itself. 

Let’s return to the MIT Technology Review report for a moment. It highlights the ‘cost paradox’: while medium-sized companies are under greater financial pressure than large ones, they still need AI to compete. For these companies, collaborating with a partner who understands their business needs and the practical aspects of the technology is ideal. This partner must facilitate regulatory compliance and data protection, simplifying procedures and ensuring the company can move forward securely and confidently. 

In fact, 98% of executives would rather sacrifice the advantage of being a pioneer than use ‘insecure and unreliable’ AI. From an SME’s perspective, this means proceeding cautiously rather than taking unnecessary risks and ensuring control over the used information is maintained. Therefore, it is essential to work with solution providers that protect the company’s data and do not reuse it without consent. In addition, incorporating practical, simple controls such as regular reviews by key employees or basic automatic systems that prevent errors or misuse will help mitigate risks.  

Another important aspect is anticipating regulatory obligations by preparing simple documentation from the outset on how the technology is used and its impact on the company. This will make it easier to adapt to future regulations. Finally, staying alert to digital security in general is key to avoiding awkward or dangerous situations and ensuring that the adoption of AI is advantageous rather than a source of concern or reputational or regulatory risk. 

In short, while 2023 and 2024 were years of curiosity and demonstrations, 2025–2026 should be a period of disciplined action. The question is not whether SMEs can afford to invest in AI, but whether they can afford not to, while their competitors are turning pilots into real productivity. 

Autor: Paul Berenguer  

Business Innovation Manager at Bové Montero y Asociados