Only 40.7% of Brazilian companies have managed to turn investments in generative artificial intelligence (GenAI) into proven financial results, according to a study by TEC.Institute in partnership with Peers Consulting + Technology, published by MIT Technology Review Brasil.
The research identified what it calls a "valley of scale" in how companies adopt the technology. Among businesses using GenAI for less than six months, 35.1% reported results above expectations. That figure fell to just 13.3% among companies with 12 to 24 months of experience with the technology, before climbing back up to 39% among organizations that have used it for more than two years.
Why returns dip after the first year
Bruno Horta, executive director at Peers, said the drop occurs during the phase when companies move past experimentation and have to reorganize their processes, systems and governance in order to scale up.
Horta said that in the early stages, companies typically see a burst of experimentation and a wave of new initiatives. As they gain maturity, he said, they need to start putting their house in order, and it is during that reorganization that the slowdown happens, before performance picks up again.
According to Horta, companies that adopted a clear strategy from the outset had greater success. He described governance as having control over what is happening: choosing which model to use, deciding whether to centralize efforts, defining what is and is not allowed, recognizing risks and discussing policy issues.
Banks outperform on business case
Among the banks surveyed, 72.7% said they had already surpassed the initial business case for their GenAI initiatives. Data governance, however, remains their biggest challenge.
Horta attributed part of that performance to the banking sector's long track record of managing crises and investing in innovation and staff training. He said AI itself is beginning to help close knowledge gaps and train professionals for new roles.
How companies are using GenAI
The study found that the leading uses of the technology are process automation, efficiency gains, improving the customer experience, and analyzing large volumes of information.
Horta said the potential of AI goes even further, describing it as a major transformation with enormous potential and saying that humans have created something very powerful.
