Artificial intelligence has become an efficiency lever for the financial ecosystem, capable of transforming days of document analysis into minutes. In the credit bureau sector, we already have proven results involving a 20% reduction in credit and fraud losses and a 66% cut in model development costs.

Experience has shown that the true competitive advantage lies not only in the use of technology, but in its structured planning. A modern and responsible AI deployment must be consolidated in 10 fundamental dimensions: regulation, strategy, technology, data, risk, governance, culture, people, social, and environment.

It is estimated that, for more consistent and effective results, institutions should adopt the 10-20-70 model: where only 10% of success comes from AI tools and 20% from IT, while a massive 70% depend on people and organizational culture. Without this focus on the human element, organizations risk falling into the “illusion of security”, a term highlighted by an IBM study, demonstrating that although 63% of financial leaders focus on regulatory risks, only 29% believe that these risks are in fact mitigated.

Effective governance requires that the AI strategy be explicitly aligned with business objectives and direct reporting to the Board. This focus on the human factor helps mitigate underestimated risks, such as the technology competency gap.

In practice, AI should be seen as an “Augmented Intelligence”, which expands human capacity instead of replacing it. According to the World Economic Forum’s “The Human Advantage” report (2026), investments in “Brain Capital” and cognitive skills are essential for AI to complement human judgment in critical decisions, which are very common in the financial sector.

Success also depends on the AI strategy being linked to business objectives, so its main uses have been: in internal tasks, internal processes, external processes and to reinvent the sector.

Once again, the human dimension is fundamental through the human intervention model adopted: Human-in-the-Loop, with mandatory human approval, is ideal for high-impact credit decisions; Human-over-the-Loop allows autonomous operation with active supervision; and Human-out-of-the-Loop, because it is fully automated, poses the greatest risk and requires the strictest governance architecture. This is a simple method that can leverage the result of AI in the risk control system in financial services.

 

 

Source: ANBC