Although AI Generative is very attractive to companies, their concrete results are still scarce, with about 95% of the projects analyzed not to achieve relevant revenue or efficiency gains.
Most Generative Artificial Intelligence (AI) projects do not achieve relevant revenue or efficiency gains in companies, reveals a study developed by Massachusetts Institute of Technology (MIT).
According to the document, although the Generative AI is very attractive to companies, Your concrete results are still scarcewith about 95% of the projects analyzed not to reach relevant gains in revenue or efficiency.
Just 5% of initiatives can speed up revenue creationbecause in most cases, projects are stagnant or residual effects on financial matters, the study points out.
In this matter, the main obstacle is not in the capacity of the models of but in the way they are applied, although many companies attribute difficulties to regulatory issues or technical limitations.
Nevertheless, the study suggests that the decisive factor is related to the integration failures of AI models in internal processes.
The application of tools, such as the It is ineffective in large organizations, as models do not automatically adapt to specific workflows or learn from companies’ internal systems.
In addition, the study also points out that more than half of the Generative AI budgets are centered on sales, while the highest return is related to the automation of internal processes, reducing costs and optimization of operations.
The study also states that companies that resort to external specialized suppliers have a success rate of about 67%, while those who choose to develop their own AI solutions internally achieve positive results in just one third of cases.
The study developed by MIT encompassed interviews to 150 executives, research with 350 employees and analysis of 300 general AI projects.
