Our work investigates how Requirements Engineering (RE) is practiced in the development of ML-enabled systems in Brazil and the challenges practitioners face in this context. ML adoption is growing worldwide, but is especially new for Brazilian companies. In 2023, 41% of these companies were adopting ML to some extent. However, the literature shows that RE can help address several challenges in engineering ML-enabled systems, despite their difficulty in handling uncertainty in such scenarios. Our research aimed to strengthen the empirical evidence on current Brazilian industrial RE practices, perceptions, and challenges in developing such systems. We ran a survey that gathered responses from 72 practitioners working on Brazilian ML projects, and conducted both quantitative (bootstrapping with confidence intervals) and qualitative (open and axial coding) analyses. Our findings provide a big picture of RE in Brazil, revealing that: (i) data scientists frequently handle RE tasks; (ii) requirements are commonly elicited through interviews and workshops; (iii) documentation often relies on interactive notebooks rather than traditional specification artifacts; (iv) and the most common RE-related challenges include poor understanding of the business domain, low customer engagement, and difficulties managing stakeholder expectations. Our first consolidated research at SBES’24 was extended into a journal article (JSERD), strengthening the evidence base and refining the insights. The extended analysis delves deep into projects’ characterization (e.g., company context, domains, tooling, ML purpose), and explicitly examines how agility influences RE-related challenges, enabling a more nuanced discussion of trade-offs and improvement opportunities for RE in ML-enabled systems. Presenting this paper at the BR Showcase @ ICSE 2026 is a unique opportunity to connect global discussions on RE for AI with the Brazilian research and industry contexts, enabling researchers and practitioners to exchange lessons learned and accelerate the adoption of more mature engineering practices.