In the last two years, artificial intelligence has definitively occupied the center of the business agenda. Boards of directors, CEOs and technology leaders discuss AI with the same intensity as they treat growth, productivity and competitiveness. Companies announced billion-dollar investments, multiplied pilot projects and accelerated the incorporation of technology in practically all areas of the business.
At first glance, it would be natural to imagine that this race was already producing an equivalent transformation within organizations. But the data shows exactly the opposite.
According to the study State of AI in the Enterprise prepared by Deloitte, in just one year, the share of workers with access to corporate AI tools went from less than 40% to around 60%. Despite this advance, only 34% of companies claim to use AI to profoundly transform their businesses, while 37% still use the technology only superficially, without changing processes or operational models. In other words, adoption accelerates; transformation, no.
However, perhaps the most emblematic data – and much discussed by current leaders – comes from MIT. The study The GenAI Divide concluded that 95% of organizations are still unable to obtain a measurable financial return from their artificial intelligence initiatives. Only 5% manage to capture consistent economic value, forming what researchers call a true “GenAI divide”: a small group turns AI into a competitive advantage, while the vast majority remain stuck in pilots, proofs of concept and incremental productivity gains.
These numbers help dismantle one of the main narratives surrounding artificial intelligence. The challenge for companies is no longer to adopt AI. Adoption is happening on a large scale. The real challenge is to transform this adoption into organizational capacity and, above all, into business results. It is precisely at this point that most projects begin to fail.
When an initiative is born from technology, the discussion usually revolves around the language model, the most sophisticated tool or the next innovation available on the market. But fundamental questions end up taking a backseat: which business problem needs to be solved? How will success be measured? What processes will be redesigned? What data will feed the models? How will this intelligence be integrated into the company’s operations?
Studies converge in showing that these questions are the real differentiator between organizations that experiment with AI and those that manage to generate value. Although AI is increasingly present in companies, 84% of organizations have not yet redesigned roles, processes or ways of working to incorporate this technology, according to the same Deloitte study. Instead of transforming the operation, most focus their efforts on increasing employees’ fluency in using existing tools. The result is predictable: lots of training, lots of pilots and little structural change.
MIT reaches the same conclusion from a different perspective. According to researchers, most companies remain dependent on generic tools that increase individual productivity, but do not learn from the business context or integrate with operational flows. More than 80% of organizations already use solutions of this type, but less than 5% are able to take customized corporate applications to production at scale. The problem, therefore, is not a lack of artificial intelligence, but a lack of integration, continuous learning and the ability to transform knowledge into operations.
The same pattern appears when analyzing the most recurrent causes of failure. Costs that grow without control, data insufficiently prepared to feed models in production and initiatives isolated from the rest of the organization repeatedly appear among the main factors that prevent value capture. These obstacles are much more related to execution than to the technology itself.
Likewise, success stories also present surprisingly similar characteristics. study drivento DellpelaEnterpriseStrategyGroupdemonstrates that organizations that structure an integrated architecture for AI can achieve ROI of up to 1,225% in four years, with a 269% return in the first year, in addition to significant reductions in operational costs, productivity gains and improvements in security and compliance.
More important than the percentage itself is what it represents: these results do not just result from the adoption of more advanced models, but from the combination of prepared infrastructure, data strategy, governance, automation and the ability to scale AI applications within the business.
Perhaps the main lesson left by the first years of the global race for AI is precisely this: the market is not facing a crisis in the adoption of artificial intelligence itself. It faces an execution crisis.
Corporate AI projects don’t fail because the technology isn’t ready yet. They fail because many organizations continue to treat artificial intelligence as a pure and simple innovation tool, when it needs to be seen as a strategic business capability. Companies that understand this difference will not necessarily be the ones that adopt AI first. They will be the ones who will be able to transform it, consistently, into productivity, competitiveness and growth over time.
*Luis Gonçalves, president of Dell Technologies for Latin America
Signed articles are the sole responsibility of the authors and do not necessarily reflect the opinion of Forbes Brasil and its editors.
Source: www.bing.com
Source link
