Co-founder, CEO at Vereda
João Souza
Raised in a farming family in Brazil; B.S. in Business Administration from the University of Southern California; lived and worked across three continents. Led over $1B in agribusiness transactions at BTG Pactual (BPAC11) and was part of the founding team of a Series B B2B marketplace in Brazil.
Questions & Answers
How is your team uniquely positioned to solve the problem you're tackling?
The founders of Vereda left São Paulo, a city of over 16 million people, to relocate 1,200 km to Rio Verde, a >100k population agricultural hub, to build from the field. We operate embedded in our customers’ day to day with speed and execution intensity that cannot be replicated remotely. We combine deep roots in agriculture with experience in tech, operating natively across both worlds — an edge that is difficult to replicate.
What problem or opportunity inspired you to start your company, and how has your vision evolved since then?
We started Vereda with the belief that farmers pay too much for inputs (seeds, fertilizers, crop protection, fuel). Living in Rio Verde, one of Brazil’s largest agricultural hubs, led us to a deeper realization: the core problem is distribution. Farmers are fragmented, hard to reach, and expensive to serve. As a result, input companies spend 10–25% of revenue on sales, while operating with single digit net margins. These costs are embedded in every transaction, making distribution the largest hidden cost in the system. Our vision evolved from reducing prices to rebuilding how supply reaches demand. Vereda aggregates farm level demand and connects it directly to manufacturers at a fraction of the cost through a digital, agent driven, data centric approach. If distribution is the bottleneck, solving it unlocks the entire agricultural value chain — from inputs to financial services.
Why are you excited about your company/product?
Vereda has the potential to fundamentally change how one of the largest industries transacts. Through a WhatsApp native AI agent, a channel with ~96% daily usage among Brazilian farmers, users buy via text, audio, or documents. The agent parses unstructured inputs, maps requests to supplier catalogs, and orchestrates multi step workflows including quote generation, credit processing, and logistics coordination. What previously required multiple human interactions is executed within a single continuous flow. Each interaction is structured into a unified farmer profile combining transaction history, behavioral signals, and external data. This enables real time offer personalization, faster decision cycles, and continuous improvement in agent performance. The result is a system where user experience and model performance compound together, making each interaction faster, more precise, and more valuable over time.
