A study with 48 thousand professionals indicates that the use of AI without training increases the risk of projects with low return

Based on global research by KPMG in 47 countries, with 48 thousand economically active people, experts in Artificial Intelligence analyze why companies start to separate initiatives that generate value from those that only scale costs Aline Lefol and Tiene Colins, co-authors of the book “IA for Business ¡ ̄ Practical guide for small and medium-sized enterprises”, this mismatch helps explain why part of AI projects does not deliver the promised financial return and, in some cases, represents an increase in operational cost.

Research shows that AI is already widely used at work, often without adequate preparation or critical evaluation of results. From an economic point of view, this gap between adoption and literacy makes it difficult to effectively capture productivity gains and increases the risk of investments with below-expected returns, especially in projects sold internally as large levers of efficiency.

In the analysis of Aline Lefol, CEO and founder of IA2YOU, a company specialized in Artificial Intelligence applied to business, the problem is not in technology, but in the way it is applied. According to the executive, initiatives that start with the tool, and not with a clear business problem, tend to generate marginal gains or rework. “IA only creates value when it is connected to measurable objectives, such as cost reduction, time or failures in specific processes. Without this, the project risks scaling expenditure faster than the result”, he evaluates.

The KPMG survey also points to challenges related to trust and governance in the use of technology, factors that contribute to a more cautious attitude of companies. After a period of strong enthusiasm, the perception grows in the market that not every application of AI justifies investment, especially in a scenario of pressure for efficiency and budget control.

For Tiene Colins, AI strategist and consultant, is, the timing reflects an important transition in adoption maturity.“The idea that AI is for any problem begins to be replaced by a more pragmatic view.When there is no structured data, integration with systems such as CRM and ERP or clear rules of operation, technology tends to amplify existing inefficiencies”, he says.

Among the most recurring failure patterns observed in different industries are the automation of still unstable processes, the absence of clear governance, unrealistic expectations of human judgment substitution, and the lack of objective metrics to track results.In these cases, AI can increase escalations to human service and operational costs, rather than reduce bottlenecks.

Strategists also warn that not every digitizable process should be automated. Activities with high regulatory or financial risk, strong dependence on empathy or absence of well-defined standards require human validation or greater control to avoid losses.

This practical, results-oriented approach is deepened in the book AI for Business & Business Practical guide for small and medium-sized businesses, authored by strategists.The work proposes objective criteria to help leaders identify which processes really justify the use of Artificial Intelligence and how to avoid projects that consume budget without generating measurable impact. In a market increasingly pressured by efficiency, the competitive differential is less by adopting AI indiscriminately and more by using it with discretion, governance and focus on real return.

About the book “IA for Business ¡ ̄” Practical Guide”

“IA for Business (Practical Guide” is the first manual aimed at small and medium-sized companies that want to transform results with Artificial Intelligence in an accessible and realistic way. Written by Aline Lefol and Tiene Colins, who guide entrepreneurs and managers in the practical application of technology in different sectors, the book presents simple tools, real cases and a strategic step by step to implement solutions without the need for technical knowledge. With clear language and practical focus, the work shows how AI is already within the reach of any entrepreneur and also offers the support of Lia, a unique virtual assistant that accompanies the reader's journey in real time.