MiFibra scales spatial analytics company-wide with CARTO AI Agents

MiFibra is a 100% fiber-optic internet provider covering 9 regions of Peru, and one of the fastest-growing ISPs in South America. It brings high-speed connectivity to underserved regions while investing in technology to double its revenue without a matching rise in costs.

MiFibra has been named Peru's fastest network for three consecutive years and was recognized as the best-rated network in South America at MWC Barcelona, results it attributes to its XGS-PON network and the commercial intelligence it runs on CARTO. It is now using CARTO to drive mass adoption in Lima, Peru's most populous department.

• Understanding market dynamics: MiFibra needed granular sociodemographic data to find underserved, high-demand areas. Manual surveys were costly and did not scale.
• Operational inefficiencies: Excel and MySQL could not automate workflows or turn location data into insight at scale.
• Customer service: Teams needed real-time monitoring of installations and network performance, which Power BI could not support for location data.
• Collections accountability: Planning and managing collections territories was essential to protect cash flow in underserved regions.

MiFibra needed a geospatial platform that could deliver live maps and analysis of network operations, automate pipelines on real-time data, enrich its own data for smarter expansion, and help it respond to the specific needs of each area it serves.

MiFibra builds churn-prediction and composite-scoring models in CARTO Workflows to prioritize expansion and flag at-risk customers. Automated pipelines refresh every 5 minutes, and CARTO Builder dashboards monitor network performance, coverage and territory planning.

• Market expansion at lower cost: Data-driven analysis replaced costly surveys, cutting market analysis costs by 50% from $2–3 per household.
• Improved collection rates: Invoice collection rose from 95% to 97% within four months.
• Optimized installation: Real-time visualization cut installation incidents by 50%.
• Competitive edge: Precise analysis of competitor coverage, speeds and expansion areas.

CARTO AI Agents in action

Nine users across sales, billing and field operations now ask questions of their data in plain language, and the agents run the queries and return maps:

• Market share and competition: Ask which zones have the fewest competitors, with results validated against MiFibra's own model.
• Power and network status: Check real-time network and power status by zone, and confirm outages before dispatching technicians.
• Customer service and call location: Map incoming service calls to surface outage clusters in real time and respond before incidents escalate.

"In a competitive telecom market, we can't afford bottlenecks. CARTO's AI Agents put instant insights into the hands of sales, operations, and expansion teams, helping us make smarter, faster decisions while keeping all of our data secure in the cloud." — Carlos Villegas, CEO & Founder, MiFibra

• Advanced geospatial features: Real-time visualization, interactive modeling and in-depth spatial analysis.
• Seamless integration: Works with MiFibra's existing tools and datasets, including Ookla data for network performance.
• Scalable platform: Cloud-native infrastructure that supports growth, predictive analytics and customer retention.
• Data enrichment: CARTO Data Observatory lets MiFibra add the sociodemographic data that replaced manual surveys.
• Rapid enablement: MiFibra's team built workflows and dashboards with minimal ramp-up, and training unlocked composite-scoring models and AI Agents.

Next, MiFibra plans to extend spatial analysis to more non-technical users through AI Agents, using them to anticipate network problems, reduce support calls and flag zones at risk of revenue loss.