The most important infrastructure for an agtech startup in Latin America isn't a cloud region or a new SDK. It's the green WhatsApp icon on a farmer's smartphone. Tecde.ai, a Medellín-based startup founded in 2024, is building its entire product on that single, ubiquitous surface. Its flagship tool, Harvis, is a digital agronomist that farmers interact with entirely through WhatsApp messages, aiming to deliver technical advice, crop predictions, and market access without requiring a new app download [tecde.ai, retrieved 2024].
The WhatsApp Wedge
For Tecde, WhatsApp is not just a convenient interface. It's a strategic wedge designed to bypass the adoption friction that has stalled countless farm management software projects. Smartphone penetration is high across Latin America, but comfort with complex, multi-screen applications is not. By operating within a familiar chat environment, Harvis can start capturing structured data from farmers immediately,reports on planting, pest sightings, irrigation, and harvest yields,building a data layer that has historically been absent for small and medium-sized operations [F6S]. This data becomes the foundation for the company's machine learning and computer vision models, which it claims can estimate raw material volumes and classify crop quality [Crunchbase]. The bet is that by entering through the chat, Tecde can eventually become the operational system for these farms.
Seed Capital and Early Traction
The company's early progress was enough to secure a $120,000 seed investment from the Techstars Chicago accelerator program, powered by J.P. Morgan, in September 2024 [CB Insights, Sep 2024]. This capital is earmarked for scaling the Harvis product and expanding its reach. Tecde reports that Harvis is already providing 24/7 technical assistance to producers, cooperatives, and agro-industries across 17 countries in Latin America [tecde.ai, retrieved 2024]. The founding team, led by CEO Luis Miguel González and CTO Juan Sebastián Ochoa, brings a focus on applied engineering, with Ochoa's background specifically in computer vision for agriculture and mining [LinkedIn, retrieved 2026]. Their public positioning emphasizes a mission to "democratize digital agriculture through technology," targeting both the farmers who need resilience and the enterprises that buy from them [tecde.ai, retrieved 2024].
| Role | Name | Focus |
|---|---|---|
| CEO | Luis Miguel González | Business strategy, market access |
| CTO | Juan Sebastián Ochoa | Computer vision, machine learning, product engineering |
The Technical Breakdown and Scale Risks
The technical architecture here inverts a typical SaaS model. Instead of a central application pulling data from edge devices, the edge is a conversational agent inside a meta-platform owned by Meta. Harvis must parse unstructured language, images, and video sent via WhatsApp, structure it into actionable agricultural data, and return timely, accurate advice. The value proposition scales with data density; more farmers using Harvis in a region improves the localized predictive models for pests, weather, and yields.
This approach introduces specific failure modes. The dependency on WhatsApp is a double-edged sword. While it enables adoption, it also means Tecde's core user experience is built atop a platform it does not control, subject to API changes, usage policies, or fees. The business model, while not publicly detailed, likely involves subscription fees from larger cooperatives or agro-industries that gain visibility into their supply chains. Proving that farmers will consistently input high-quality data for long enough to train reliable models is an unproven hurdle. Furthermore, the path from a helpful chatbot to a mission-critical system that commands significant enterprise value is long and requires deep domain integration that goes far beyond conversational interface.
The Path to a Data Monopoly
Tecde's ambition appears to be the creation of a proprietary agricultural data network. If Harvis becomes the default way smallholders manage their operations, the company would own a unique dataset on micro-climates, soil health, and production cycles across Latin America. This data asset could be immensely valuable for commodity forecasting, insurance underwriting, and input manufacturing. The next twelve months will be about proving that the wedge works at a commercial scale. Key signals to watch include the announcement of paid pilots with named cooperatives, the publication of any case studies showing yield improvements or cost savings, and a subsequent funding round to build out a full-stack team beyond the core founding duo. The risk is that the product remains a useful but peripheral advisory tool, never capturing the comprehensive data or trust required to become foundational infrastructure.
Sources
- [tecde.ai, retrieved 2024] Tecde · Democratizamos la agricultura digital a través de tecnología | https://www.tecde.ai
- [F6S] Tecde.ai | https://www.f6s.com/company/tecde.ai
- [Crunchbase] Tecde Company Profile | https://www.crunchbase.com
- [CB Insights, Sep 2024] Tecde.ai Financials | https://www.cbinsights.com/company/tecde/financials
- [LinkedIn, retrieved 2026] Juan Sebastian Ochoa - CTO - Tecde Inc. | https://www.linkedin.com/in/juanochoa-ctotecde/
- [LinkedIn, retrieved 2026] Luis Miguel González - CEO - Tecde.ai | https://co.linkedin.com/in/luisgonzalez-ceotecde/en