A hands-on AI adoption workshop showing how non-technical Mercado Libre teams can enable Claude Cowork to analyze seller data, generate reports, and build reusable workflows.

AI Fluency Is Becoming an Enterprise Advantage

Enterprise AI adoption is no longer limited to technical teams. As organizations move from experimentation to everyday use, the real challenge is helping business teams apply AI to the work they already do: analyzing information, preparing reports, identifying risks, and making faster decisions.

Mercado Libre, Latin America’s leading e-commerce and fintech ecosystem, has been driving a strong internal push to help collaborators adopt AI tools across their daily workflows. For teams supporting sellers, that meant finding a tool that could work with real business data, reduce manual effort and be useful without requiring a technical background.

That context made Claude Cowork a strong fit.

The Challenge: Bringing AI Into Seller Support Workflows

Mercado Libre’s seller support teams work close to the operational reality of the marketplace. They review seller performance, identify risk signals, escalate critical cases, and prepare commercial updates for internal stakeholders.

The work is analytical, high-volume, and often repetitive. But the users are not engineers or data scientists. They needed a way to interact with data, summarize findings, and create reusable workflows without depending on technical teams for every task.

The goal of the workshop was not to teach AI theory. It was to help participants use Claude Cowork to solve the kinds of problems they already face in their day-to-day work.

Why Claude Cowork Was the Right Fit

Claude Cowork was selected because it matched the needs of this audience: a tool designed to work alongside users on real files, projects, and workflows.

For Mercado Libre’s seller support teams, the most relevant capabilities were its ability to preserve context, connect with external tools and data sources through MCP, and turn repeatable tasks into reusable Skills. Together, these features made it possible to move beyond one-off prompts and start building practical workflows for recurring business tasks.

Inside the Claude Cowork Workshop

Marviks team led a hands-on Claude Cowork workshop at Mercado Libre’s offices.

The session was designed for non-technical users from the start. Instead of focusing on model architecture or technical theory, the workshop centered on practical exercises using seller-related data and familiar workflows.

Within the first part of the session, participants were already working in Claude Cowork, testing how it could help them analyze seller files, structure reports, identify risk signals, and save recurring workflows as Skills.

The goal wasn't to teach people about AI. It was to make sure that when they walked out of the room, they had a tool that would save them time the very next morning.

What Claude Cowork Is Built On

Before the exercises, the Marvik team walked through the three pillars that shape how Claude Cowork behaves for business users.

1. ‍Memory. Claude Cowork retains context between sessions. It learns each user's role, their usual tasks, their preferred formats. Over time, it stops feeling like a new tool you have to re-explain yourself to every morning.‍
2. MCP Connectors (Model Context Protocol). Rather than waiting for data to be pasted into a chat window, Claude Cowork can connect directly to external tools and data sources. For a team that spends significant time exporting and reformatting files, this matters.‍
3. Skills. Skills are saved, reusable workflows that run on demand. One analyst builds a risk detection workflow once, saves it as a Skill, and the whole team can use it.

What Participants Built in Four Hours

Business operators, none with a technical background, completed five exercises using live commercial data in four hours. Each left with at least one working Skill configured to their own workflow.

That's the version of AI adoption that actually sticks: participants solving their own problems with their own data, not watching demos of someone else's use case.

From AI Training to Daily Use

The workshop showed that enterprise AI adoption does not depend only on deploying more tools. It depends on giving the right teams the right context, examples, and workflows to make those tools useful.

For Mercado Libre’s seller support teams, Claude Cowork became a practical way to apply AI to real operational work: analyzing data, preparing reports, identifying priorities, and reducing repetitive manual steps.

For Marvik and Anthropic, the case reinforces a broader point: AI adoption scales when business users can work with AI in the same environment where their daily work already happens.