WhatsApp-native AI for customer service, sales, and collections at scale
Mundiale operates a conversational AI platform purpose-built for WhatsApp, handling 100M+ messages monthly across customer service, sales, and debt collection workflows. The tech stack reveals a data-science-forward org: Python + scikit-learn + Transformers + RAG + BERT sit alongside Node.js and MongoDB, with BI tools (Power BI, Looker, Tableau) wired in—suggesting heavy emphasis on funnel optimization and agent performance tracking. Active hiring across product, data, and ops (versus engineering), combined with projects centered on conversion optimization, prompt engineering, and reducing human handoffs, indicates the core challenge is converting high-volume conversations into revenue while maintaining bot reliability.
Mundiale is a Brazilian AI company founded in 2001, headquartered in Belo Horizonte, with 500+ employees. The platform automates customer interactions—inbound service inquiries, outbound sales, and collections—exclusively via WhatsApp, the dominant messaging layer in Latin America. The business operates at meaningful scale: over 100M messages exchanged monthly across hundreds of client companies. The product roadmap prioritizes chatbot reliability (bot uptime is a tracked pain point), conversion tuning through A/B testing, and reducing the need for human agent escalation, which directly impacts COGS and gross margin per conversation.
Node.js, MongoDB, Python, RabbitMQ, Git, Docker for backend. Data: scikit-learn, Transformers, BERT, RAG. Analytics: Power BI, Looker, Tableau. Operations: HubSpot, Scrum, Kanban.
New chatbot development, conversion optimization, A/B testing, prompt engineering + RAG strategy, campaign management (email/push/SMS), and reducing human handoffs to improve automation efficiency.
Other companies in the same industry, closest in size