Naspers and Prosus Pivot to Agentic AI with Zapia Investment and Scaled Internal Deployments

Global consumer internet group Naspers and its international investment arm Prosus are driving a major shift toward agentic AI, moving beyond conversational tools to autonomous systems that execute real-world tasks. As part of this transition, Prosus has thrown its backing behind Zapia, an action-oriented personal AI agent developed by BrainLogic AI, while Naspers has shared key operational insights from its own massive internal deployments of generative AI tools across its global network.

Inside Zapia and the Rise of Action-Oriented AI

Zapia represents a distinct leap from traditional search-and-answer assistants to autonomous agentic platforms. Available on the Google Play Store with a 4.3-star rating from over 43,800 reviews, Zapia is designed to perform complex digital chores without manual setup. Users interact with the agent simply by chatting or sending voice notes.

The platform integrates deeply with daily communication and productivity tools. Through WhatsApp, Zapia can automatically read, summarize, and respond to messages, transcribe voice notes, and contact external businesses or restaurants. It also syncs with Google Calendar and Gmail to manage schedules, set recurring reminders, and draft emails. For commerce and travel, Zapia automatically compares prices across marketplaces like Amazon and MercadoLibre, tracks products for price drops, searches flights via Skyscanner and Despegar, and books restaurant tables through platforms like OpenTable, Meitre, and AgendaPro. Under the hood, Zapia utilizes a suite of advanced AI models depending on the task, including GPT, Claude, Claude Cowork, Claude Code, Gemini, Sonnet, Opus, and OpenClaw, which is also referred to as Toqanclaw.

Naspers Shares Blueprint for Enterprise AI Scale

As Prosus backs consumer-facing agents like Zapia, its parent group Naspers is drawing on years of internal AI implementation. Serving a quarter of the global population across more than 100 countries, Naspers has long relied on machine learning, keeping hundreds of predictive models in active production to manage its vast transaction volumes.

Naspers began experimenting with Large Language Models in 2019, utilizing early systems like BERT and GPT-2 to process unstructured data. Between 2020 and 2021, the group conducted more than 20 practical field tests, exploring use cases from code automation and bug fixing to document synthesis and educational Q&A. In the summer of 2022, Naspers launched a dedicated personal AI assistant for its own workforce. Today, this internal generative AI tool is actively used by approximately 13,000 colleagues across 24 group companies, providing the organization with a massive testing ground for scalable enterprise AI.

Distinguishing Zapia from Offline Sharing Tool Zapya

While the AI ecosystem focuses on the agentic capabilities of BrainLogic AI's Zapia, the name should not be confused with Zapya, a completely separate peer-to-peer file-sharing application. Developed by DewMobile Inc. and initially conceived in Silicon Valley in 2012 as Kuai Ya for the Chinese market, Zapya operates as an offline utility.

Zapya gained massive traction in regions with poor internet infrastructure and low connectivity, such as Myanmar and Pakistan, by allowing users to transfer files directly between devices without cellular data or internet access. Although both tools address efficiency, Zapya relies on local peer-to-peer transfer protocols similar to Bluetooth, whereas the newly highlighted Zapia relies on advanced cloud-based LLMs to automate complex digital workflows.

What this means for Africa: As South African internet giant Naspers scales its internal artificial intelligence systems, its backing of WhatsApp-compatible agents like Zapia highlights a clear path for highly accessible, chat-driven automation across African businesses.