What made the challenge harder was organizational structure. The provider had more than 13 separate churn-reduction programs running in parallel, each owned by a different business unit, each operating on different data, different models, and different assumptions about what was driving customers to leave. There was no shared framework, no common methodology, and no way for the Consumer, Business, and Network segments to coordinate action. Teams were working from intuition as often as from data. Individual programs generated insights that rarely reached the people or functions best positioned to act on them. The result was a fragmented retention effort that consumed significant resources without producing the coherent, enterprise-wide impact the business needed. The provider recognized that fixing churn meant fixing how the organization understood, measured, and responded to it.
Solution
Our team at Brillio approached the engagement in structured phases, starting with the scientific foundation before moving to execution. The first phase deployed Causation Intelligence through a Tech Acceleration Engine that automated AI science workflows and introduced self-service, governed business intelligence at scale. We built causal machine learning models to isolate high-risk customer deciles, identify the highest churn-prone segments, and trace root causes across three distinct driver categories: customer service behavior, network performance, and value-based switching. This shifted the operating model from intuition-led decisions to model-driven action.
The second phase moved from prediction to intervention. We deployed specialized Agentic AI pods targeting the highest-impact use cases across the business. In the Consumer segment, the pods automated ticket triage and resolution, enabled proactive order monitoring with auto-resolution, streamlined billing inquiry handling, and provided call center support for network-related queries. In the Business segment, agentic processes streamlined the opportunity-to-order journey and reduced operating expenses. In the Network segment, intelligent diagnostics enabled outage analysis and automated recommendations, while Agentic Lease Automation optimized vendor and operational workflows.
Underpinning all of this was a centralized Agentic Marketplace and Control Tower that gave every business unit a single view of all AI agents and their performance insights. The program also delivered 32 or more reusable churn-focused AI assets built natively within the client’s ecosystem, alongside structured enablement that included webinars for approximately 500 employees and agentic AI workshops for approximately 300, ensuring the organization could sustain and build on what was deployed.
