Azercell Builds a Unified, AI-Driven Data Platform with Starburst Enterprise

About

Azercell is the largest telecommunications operator in Azerbaijan, operating at a massive data scale across network, customer, and business systems. As a long-established incumbent, the company manages large volumes of data across multiple legacy systems and silos. Sabir Mardanov, Chief Data and AI Officer, leads Azercell’s efforts to monetize data and AI—both internally through analytics and externally through new data-driven services.

Challenge

Azercell’s data landscape consisted of multiple siloed systems, with data arriving at different speeds, in different formats, and at different frequencies. As the company moved toward a more data-driven culture, it needed to:

• Enable self-service analytics across decentralized business teams
• Provide unified access to data across data lakes, warehouses, and data marts
• Enforce robust governance and access controls in a regulated environment
• Reduce reliance on centralized data teams and manual data requests
• Support emerging AI-driven use cases and new ways of interacting with data

The organization was also transitioning away from legacy query engines such as the open-source Trino and Impala, seeking a more scalable, unified solution.

 

Solution

Azercell deployed Starburst Enterprise as a federated query engine to unify access across its distributed data ecosystem. Starburst acts as a virtualization layer across data lakes, warehouses, and data marts, enabling teams to query distributed data without movement.

The platform supports Azercell’s self-service data strategy, allowing decentralized teams—such as finance, network, and customer intelligence—to create and consume data products. Starburst also provides centralized governance through role-based access controls, with plans to expand to attribute-based access control (ABAC).

In parallel, Azercell integrates Starburst with its internal AI ecosystem. By connecting Starburst to AWS Bedrock and its in-house generative AI platform, the company enables capabilities such as query acceleration, metadata exploration, and AI-driven data agents.

Results

• Unified access across 40+ data sources: Starburst enables federated queries across Azercell’s distributed data ecosystem without requiring data movement.
• Expanded self-service analytics: Decentralized business teams can now directly access and analyze data, reducing reliance on centralized data teams.
• 25% reduction in support requests: Within the first quarter of completing deployment and enterprise rollout, Azercell saw a 25% reduction in data-related support tickets as more users gained self-service capabilities.
• 5 production AI use cases launched: Azercell has deployed approximately five production AI use cases—including AI agents and agentic workflows like the AI Board Navigator, a C-suite decision-support tool—leveraging Starburst MCP and Starburst Data products.
• Simplified data product development: Teams can create, share, and consume governed data products across the organization.
• More flexible data integration: Starburst enables lighter, more flexible data pipelines by reducing the need for traditional ETL and supporting near real-time data access.

With Starburst Enterprise as a unified data access layer, Azercell is building a scalable foundation for both self-service analytics and AI-driven innovation across the business.