Agentic customer operations: End-to-end agentic customer lifecycle

Agentic customer operations: End-to-end agentic customer lifecycle reimagines how CSPs manage the customer journey by connecting BSS, customer care and network operations within a single agentic ecosystem. Instead of customers experiencing fragmented processes and repeated handoffs, 19 purpose-built AI agents collaborate across the lifecycle, from onboarding and activation through billing, care and service assurance, carrying customer context across traditional operational boundaries.

The innovation lies in enabling AI agents to work together as one coordinated customer operation. Agents follow an Intake, Diagnose, Act and Confirm model to resolve order fallout, address billing issues, support upsell and retention opportunities, and detect and remediate service problems before customers need to make contact. A supervised autonomy model provides guardrails and auditability, escalating higher-risk or judgement-based decisions to people while enabling routine activities to be handled autonomously.

For CSPs, this approach has the potential to tackle significant sources of cost and customer frustration, including stuck orders, revenue leakage, repeat contacts, lengthy resolution times and reliance on specialist teams. Proposed targets include a 30% reduction in order-fallout detection and resolution time, at least 50% of in-scope order exceptions resolved without manual intervention, a 30% reduction in customer-care handling and resolution time, and 40% fewer human handoffs and escalations. By sharing context across domains, the platform also creates opportunities to improve first-contact resolution, proactively communicate service issues and turn customer interactions into retention or revenue opportunities.

Built around TM Forum Open Digital Architecture (ODA), Open APIs and Autonomous Operations principles, the Catalyst is designed to work incrementally with existing CSP environments rather than becoming another isolated point solution. A governed data foundation and closed-loop learning allow every resolution to strengthen future agent decision-making, while all actions remain traceable. Ultimately, the project aims to demonstrate how CSPs can deliver a more seamless and individualised customer experience while lowering cost to serve, protecting revenue and progressing towards scalable, autonomous customer operations