From dashboards to autonomous intelligence across 18 countries
Mercado Libre is not just Latin America’s largest e-commerce platform. It is one of the most complex digital ecosystems in the region. Operating across 18 countries, from Mexico to Argentina, the company processes hundreds of billions of behavioral events each year. Every click, search, purchase, shipment, and delivery creates data.
The ambition was never just to analyze that data. It was to make it actionable for everyone.
Mercado Libre wanted every team, across product, logistics, marketing, and operations, to make fast, confident decisions grounded in real behavior. Not after a report was requested. Not after a weekly meeting. In real time.
That required more than dashboards. It required a new operating model.
The Reality of Analytics at Continental Scale
Running analytics for a company of Mercado Libre’s size introduces challenges most organizations never encounter. Logistics stretch across dense urban centers and remote rural regions. Each market has its own infrastructure, partners, and regulatory environment. What works in São Paulo may not work in Córdoba or Mexico City.
Mercado Libre had already built a strong analytics culture. In a single year, access to Amplitude expanded to more than 10,000 employees, with over 2,000 monthly active users. The company was processing 782 billion events.
But scale creates friction.
When thousands of people use analytics, consistency becomes harder. Knowledge fragments across teams. Analysts become bottlenecks. Insights depend on who asks the right question at the right time. Even when problems are identified, turning insight into coordinated action across functions requires effort that doesn’t scale.
The question shifted from “How do we analyze more?” to “How do we reduce the distance between insight and action?”
Building an AI-Native Operating Model
Together, we approached the transformation in layers.
First, we strengthened the behavioral data foundation. Amplitude became the governed source of truth for customer journeys, capturing the full path from initial interaction to final delivery. The priority was clarity and trust. Without shared context, automation has no anchor.
Then we introduced automation directly into decision workflows.
Instead of dashboards waiting to be checked, AI agents began actively monitoring performance. They detected anomalies, summarized trends, and distributed insights automatically. Teams no longer had to remember to look for problems. The system surfaced them.
Adoption moved quickly. Within months, dozens of users had built more than a hundred agents. Most of those agents were not just generating reports, but triggering meaningful follow-up actions. Emails generated by agents saw open rates above 50%, a signal that the insights were not noise. They were relevant and timely.
Finally, intelligence was distributed into the tools employees already used. Through Model Context Protocol, Mercado Libre teams could query governed Amplitude data directly from AI tools like Claude or their internal assistant. Instead of navigating complex interfaces, employees could ask natural language questions grounded in real behavioral data.
Analytics stopped being a destination. It became infrastructure.
A Concrete Example: Flex Shipments in Brazil
One of the clearest demonstrations of this shift happened within Mercado Libre’s Flex delivery program in Brazil.
Flex allows sellers to choose independent couriers capable of meeting strict service-level agreements. But when sellers searched for couriers and found none, friction followed. Failed searches led to slower deliveries and growing dissatisfaction.
Previously, identifying coverage gaps required manual analysis. Teams often reacted after issues became visible in performance metrics or customer complaints. Prioritization was influenced by anecdotes as much as data.
We worked with Mercado Libre teams to create a dashboard that made unmet courier demand visible at a granular geographic level. For the first time, product, transport, and business teams were aligned around the same real-time view of where sellers were searching and coming up empty.
That visibility changed the conversation. Instead of debating where to focus, teams could quantify impact.
The real transformation came when AI agents began monitoring that dashboard continuously. Rather than waiting for someone to check performance, agents automatically surfaced regressions or emerging gaps. This created a closed loop. Teams identified high-impact zones, onboarded couriers with stronger SLA performance, monitored improvements, and refined strategy as demand shifted.
The results were measurable. Failed courier searches dropped by nearly ten percent. 60% of Flex sellers in Brazil now complete shipments through the feature. SLA performance reached 98%, outperforming shipments outside the program.
More importantly, this was not a one-time fix. It became an ongoing optimization system.
Our Role
Technology enabled the shift, but operationalizing it required structure.
Minders embedded into Mercado Libre’s cadence through executive reviews, enablement sessions, and cross-functional workshops. We helped translate business priorities into concrete analytics initiatives. We supported governance, experimentation, marketing analytics expansion, and AI rollout.
At this scale, transformation is not about installing a tool. It is about aligning people, processes, and data so intelligence can flow continuously across the organization.
Our role was to bridge that gap between capability and adoption.
From Insight to Continuous Action
Mercado Libre’s journey toward becoming AI-native continues to evolve. The next step is deeper integration, where internal agents can trigger actions directly from workflow tools, whether that means creating cohorts, launching experiments, or opening governance tickets.
The objective is simple: reduce the time between sensing, deciding, and acting until analytics becomes invisible. Not because it is unimportant, but because it is embedded into how work gets done.
For large enterprises operating across multiple markets, this is the future of competitive advantage.
Not more reports.
Not more dashboards.
Smarter systems that move as fast as the business.
Ready to Operationalize AI in Your Organization?
Most enterprises already have data. Many have dashboards. Few have turned analytics into a continuous decision engine.
If your teams still rely on manual analysis, delayed insights, or siloed reporting, the opportunity is clear.
Minders helps organizations move from analytics adoption to operational intelligence. We design the governance, workflows, and enablement models that make AI practical at scale.
