Fabricio Bloisi

AI Serving 1 Billion Customers

The core message of the session was that Prosus sees AI as a near-term operating shift for its businesses, not only as a customer-facing chatbot layer. The speaker opened Prosus Forward by framing the event around innovation and disruption “made in Europe,” Latin America, and India, then grounded the opportunity in Prosus’s current scale: one billion customers and 5,000,000 partners served through apps across delivery, fintech, events, and other services. The key value proposition was that, over the next one to three years, Prosus expects much more AI in its services, products, assistance experiences, and internal company operations. The speaker then explained the AI flywheel Prosus believes companies can run today. At the foundation is company data: while many LLMs are trained on internet data, Prosus believes internal data that captures what customers want, buy, and avoid will be the basis for companies positioned to win in the AI revolution. The second layer is intelligence embedded across company operations rather than confined to chatbots. The speaker described a future where every meeting, discussion, and internal connection uses AI to help the company operate better. This intelligence, combined with customer data, is expected to power apps that understand customers more effectively. Prosus’s concrete example was its large commerce model, described as trained on 500,000,000 daily interactions to understand what users are looking for, what they want, what they bought, and what they do not want to buy. The speaker said this model was already running and had been discussed three months earlier, then moved to what would run on top of it. OpenCLO was presented as a recent influence that changed how Prosus thinks AI can help: it showed AI can be smarter, build software and tools, understand intent, and actively help users. At the same time, the speaker stressed that OpenCLO is difficult to use, hard to set up, unsafe, and unpredictable, making it unsuitable to run at scale today. The concluding announcement was Prosus’s TokenClaw approach, described as something like OpenCLO but running in the cloud, easier to set up, using company data, and already operating inside Prosus. The speaker said employees across finance, legal, HR, tech, and AI are creating apps, self-learning loops, and tools that capture meeting data and company knowledge to make teams faster and smarter. Finally, Prosus plans to extend TokenClaw beyond internal use to its 5,000,000 partners, including restaurants, and then through Zapier to offer OpenCLO-like features to its one billion customers. The session ended with excitement about creating this AI future from Europe and sharing lessons from Prosus’s recent experience.