AI adoption resistance slows Brazilian productivity gains
AI adoption resistance, not the technology itself, is slowing gains at Brazilian firms. Here is what investors should watch.

Magazine Luiza and CI&T both encountered significant employee adoption challenges. Their experience points to a broader execution gap across Brazilian corporates. The obstacle is not capability. It is behaviour.
Frederico Trajano, Magazine Luiza’s chief executive, said the retailer employs around 2,000 engineers. That headcount is unlikely to grow in the near term, he added. Properly implemented, AI raises productivity and reduces pressure to hire. However, the shift demands changes in how employees work and how managers set expectations.
Some staff were reluctant to disclose their AI use. Others feared the technology would replace them. Trajano said companies must reward employees who learn to use AI effectively. That cultural shift, he argued, matters as much as the software itself.
Why are employees resisting AI tools?
CI&T’s experience reinforces that view. Vinicius Martinelli, a vice-president at the technology services firm, said CI&T built its own AI development platform roughly three years ago. Employees initially ignored it, despite the company’s investment. A structured programme was required to lift adoption rates and address fears over job security.
The message for investors is direct: AI spending does not guarantee operational gains. Returns depend on whether organisations can change behaviour at scale.
Martinelli said AI use is now widespread at CI&T, with strong productivity improvements in software development. However, both executives flagged a risk they call workslop. The term describes AI-generated output that looks polished but lacks genuine analysis. A lengthy presentation may contain generic conclusions and poorly understood data. Presenters sometimes fail to review the figures behind their own materials.
Brazilian finance departments lag on AI maturity
That risk shifts value firmly towards human judgement. When AI makes information cheap to produce, managers must decide where investment creates measurable returns. Companies therefore need training, controls and clear accountability — not just access to tools.
Survey data adds context. Research by Barte, conducted among more than 100 CFOs and financial leaders at medium and large Brazilian companies, found that 89% of finance departments sit within the first two levels of a five-level AI-maturity scale. No company surveyed reached either of the two highest levels, according to Barte’s findings.
Meanwhile, Martinelli argues the pilot phase is largely over. Companies should now identify specific processes where AI can deliver material improvements. That changes the investment question entirely. The issue is no longer whether AI works. It is whether organisations can absorb it quickly enough to matter.
What should investors watch next?
For listed companies, AI adoption resistance may affect hiring plans, operating costs and returns on technology spending. Staffing trends in engineering and technology functions are worth monitoring closely. So are software productivity metrics and evidence of financial returns from AI programmes.
The strongest performers will combine capable tools with disciplined human oversight. Training investment and clear accountability structures will separate genuine adopters from those accumulating expensive pilots. Investors should track both the pace of adoption and the quality of the outcomes Brazilian companies can demonstrate as they scale AI across their operations.
Quick answers
Employees at firms such as Magazine Luiza and CI&T resisted AI tools due to fears over job security and reluctance to disclose AI use. Structured programmes addressing both behaviour change and accountability were required to lift adoption rates.
Barte research among more than 100 CFOs and financial leaders at medium and large Brazilian companies found that 89% of finance departments sit within the first two levels of a five-level AI-maturity scale, with no company reaching the top two levels.
Workslop is a term used by CI&T to describe AI-generated output that appears polished but lacks genuine analysis or well-understood data. It signals that AI investment requires human oversight and accountability structures to deliver real business returns.




Comments