OpsSights

a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk
a computer monitor sitting on top of a wooden desk

What we built: From endless tickets to intelligent answers in one click

An overloaded support team, thousands of repetitive questions, and processes that did not scale.

In this context, we built OpsSights, a multi-agent system capable of interpreting metrics, explaining indicators, and resolving operational questions without human intervention.

A POC that showed how AI can become the new first line of support, fast, accurate, and reliable.

The challenge: When questions grow faster than the team can answer

Data support was stuck in a reactive loop. Analysts, managers, executives, and franchisees constantly raised questions about core metrics (ACR, SOS, Training, REV), missing data, and operational indicators.

This resulted in:

  • Overloaded support channels

  • Poorly structured tickets

  • Endless back-and-forth between support, development, and stakeholders

  • Slow resolution of simple issues

  • Inefficient use of engineering time

The mission: bring intelligence to support operations without adding load to the team.

How we built it: A multi-agent assistant that understands metrics, business logic, and SQL

An AI-powered support agent with a multi-agent architecture, trained on the company’s internal KPI definitions, business rules, and technical processes.

Key capabilities:

  • Natural language interpretation of metric-related questions

  • Contextual explanations of calculations, data sources, and variations

  • Dynamic SQL generation for technical support

  • Standardized, consistent, and reliable responses

At the core of the solution is a context router that directs each query to the appropriate agent, business, technical, or training, reducing noise, ambiguity, and information overload.

The guiding principle was simple: a support system that thinks, understands, and responds like a senior analyst, in seconds.

Key accomplishments that we're proud of:

  • 63% fewer tickets generated

  • 97% reduction in back-and-forth between stakeholders, support, and development

  • 100% improvement in ticket quality and structure

  • 55% estimated savings in engineering costs

The team

Mauricio Bergallo
Mauricio Bergallo

Mauricio Bergallo

Data Engineer

Jose Ruiz
Jose Ruiz

Jose Ruiz

Data Engineer

Raul Ubillos
Raul Ubillos

Raul Ubillos

Developer

Julian de Tomasi
Julian de Tomasi

Julian de Tomasi

Data Engineer

Rodrigo Arana
Rodrigo Arana

Rodrigo Arana

Data Engineer

Julian Jarazo
Julian Jarazo

Julian Jarazo

Developer

Industries

AI Agents

Foodtech

Reskilling for the Future

Learning turned into action

We are a community driven by curiosity and collaboration. Reskilling is part of our everyday practice because change happens through us. The future is made by believers, and the revolution is human.

Powered by

© 2025

Powered by

© 2025

Powered by

© 2025