Case study · Agribusiness · Guatemala · Client since 2018

Grupo Entre Ríos

52% → 85%

capacity used per truck

An AI model plans weekly raw-material pickup based on production, weather and demand. We also built their intake, budgeting and maintenance systems.

Capability: Build

Grupo Entre Ríos

Executive summary

One of Guatemala’s three largest rubber exporters went from using 52% of its pickup-truck capacity to at least 85%, with an AI model that plans weekly collection from production, weather and demand — built on top of the operational platform Ciancoders already maintained for the group.

The story

Client context: Grupo Entre Ríos is one of Guatemala’s three largest rubber exporters. Its operation depends on collecting raw material from dozens of points every week. A Ciancoders client since 2018.
Challenge: Pickup trucks left using only 52% of their capacity on average. Every partial trip is a logistics cost repeated week after week, in an operation where each point’s output varies with weather and demand.
Insight: A planning model is only useful if the operation feeding it is recorded reliably. That is why the work started with the operational foundation — rubber intake, budgeting and maintenance — and not with the model.
Approach: First, the internal platform that records daily operations and the group’s corporate website. Then, with that foundation in production, an AI model integrated into weekly planning and evaluated against the baseline truck utilization.
What we built: The internal platform for rubber intake, budgeting and maintenance; the corporate website; and the AI model that generates the weekly pickup plan.
Engineering decisions: Three inputs — production, weather and demand — and one operational output: the weekly pickup plan. The model produces a plan the logistics team executes; the result is measured with the metric the business cares about, capacity used per truck.
How Cian-OS applies: Intent: the success metric was truck utilization, not model accuracy. Foundation: the operational platform came before the model. Build: the model was integrated into the existing weekly flow. Verify: the result was compared against the 52% baseline. Operate: the systems keep evolving with the same team. (A retrospective reading: Cian-OS formalizes patterns learned on projects like this one.)
Business outcome: Average truck utilization rose from 52% to at least 85%.
What happened next: The intake, budgeting and maintenance systems keep evolving with the same Ciancoders team.