According to the MIT report on the AI status in the company in 2025 The Genai divides: State of Ai in Business 2025, of which Fortune has been made, despite the fact that The generative AI is promising for companies, most initiatives With this technology that are launched in companies in the United States to get rapid income from use They are failing.
In companies there is some urgency to integrate new models into their processes, and there are 5% of AI pilots who achieve an acceleration in income. But as we have mentioned, most pilots end up stopping, and do not get practically impact on results accounts.
For the elaboration of the study, 150 interviews with managers have been carried out, as well as a survey between 350 employees and an analysis of 300 public deployments of AI. Their results mark a clear dividing line between the projects that are successful and those that fail.
American companies, to date, have invested between 35,000 and 45,000 million dollars in generative projects, but most of the efforts they have made have not gone from the pilot phase, and have not had any impact, or this has been very scarce.
The main problem, apparently, has nothing to do with the quality of the models that are used in the projects. Its failure is more related to the lack of integration, learning and alignment with corporate workflows of companies.
Companies also focus on projects in which generative the pilots may not offer the best results. Thus, it is common for them to invest in solutions. For marketing and sales, but the best results are achieved by the projects for the automation of the backoffice, as well as for the streamlining of internal processes.
On the other hand, according to the report, companies that achieve success in their generative AI pilots tend to use specialized solutions, as well as to seek agreements with third parties for its implementation. As for the projects developed internally, they fail much more frequently. This is leading to a fall in confidence in the initiatives related to AI among those responsible for not a few companies.
By sectors, those that have achieved smaller impacts are those of professional services, health and pharmacy, as well as consumption and retail, financial services, and energy and materials. Many companies are having difficulty quantifying the benefits of their IA deployments, but 80% of sectors such as technological and media, which have achieved more impacts, already expect a reduction in their hiring thanks to AI throughout the next two years.
Despite this data, the impact of the AI in general, and of the generative in particular, in the templates of the companies is variable. The employee cuts that entail mostly affect areas that are not central, as well as externalized positions. In the latter case, specifically, it is estimated that. Between 5% and 20% of externalized positions have been affected by the deployment of AI.
The study has also revealed that the workers prefer generic tools, such as Chatgpt, to the specialized ones. Even if they use the same models. In particular, the familiarity and flexibility of Chatgpt has grow of adoption in the shadow of AI in you.
That is why the study indicates that companies must assess the needs of their workers and adapt their policies according to the security problems they may raise, instead of prohibiting them completely. As for company tools with generative, with higher costs, they are perceived by employees as more rigid and less effective.
Facing the future, the study states that Companies should consider simpler strategies for the adoption of AI and the generative AI. Thus, instead of developing complex proprietary systems, they may obtain better results if they adapt tools already available to meet company policies. In addition, doing this can also reduce the amount of hours of training in which workers may need.
