94% of HR professionals already use AI daily, but only 1 in 5 trusts it for sensitive decisions
A survey by LG lugar de gente with FIA Business School, conducted during CONARH 2026 among more than 1,100 HR professionals from across Brazil, shows Artificial Intelligence already saves an average of 5.4 working hours per week, but still occupies a small space in the more delicate decisions.
In focus: LG lugar de gente, in partnership with FIA Business School, ran a field survey during CONARH 2026, Brazil's largest HR congress, with 1,130 HR professionals from 26 states and more than 40 industries. The result, released in September: 94% of respondents already use Artificial Intelligence in their day-to-day work, saving an average of 5.4 hours per week, equivalent to about 12% of the workweek. The most common uses are screening and recruitment (39%), internal-service chatbots (38%) and text and presentation generation (31%).
The figure that draws the most attention, however, is what's left out: only 21% of professionals use AI to support decisions, and 20% to predict turnover. In other words, AI is already deeply embedded in HR's operational and repetitive tasks, but human judgment still concentrates the most sensitive decisions, such as performance evaluation, succession and termination.
This is exactly the pattern we've been observing: AI adoption in HR moves fast on the operational side (screening, service, content generation) and moves slowly, with good reason, on the decision-making side, where a mistake costs more in labor, LGPD and reputational terms. The gap between these two speeds is, today, the main point of attention for whoever designs the company's policy on AI use in HR.
We help companies design their internal policy on AI use in HR and personnel administration, clearly separating what can be safely automated from what still requires a documented human decision, with special attention to recruitment, evaluation and termination. Learn about Nexus's support for law firms and HR teams.
Does your company already have a written policy on where AI can decide on its own in HR, and where it can only support a human decision?