Bain: AI to greatly increase network operator expenses; network re-engineering needed!

Bain & Company said AI agents planned for use across network operations and other functions could increase the operating-cost burden for telecom operators if legacy processes and supporting structures were not retired alongside automation. The issue ties to plans for autonomous networks and whether added agent capabilities reduce overall workflow costs rather than adding incremental expenses.

The company’s assessment cited TM Forum reporting in June that 81% of 80 surveyed operators targeted Level 4 autonomous networks or higher by 2030, with 20% expecting to reach the threshold by 2027. TM Forum’s Autonomous Networks framework described Level 4 as moving beyond rule-based automation toward closed-loop, intent-driven decision-making within defined network domains, including production deployments, agent-based operating architectures, and effectiveness metrics.

Bain estimated that AI agents and associated token consumption could represent 20% to 30% of a telecom operator’s operating-cost base within the next three to five years, leaving conventional operating costs at 70% to 80%. Bain said the risk comes from creating a parallel operating model by layering agentic AI onto existing network-operations processes, so expenses for AI compute and orchestration rise while legacy operating centers, monitoring platforms, software licenses, managed services, and manual handoffs remain in place.

Bain recommended network re-engineering instead of task-level automation, including reducing manual monitoring and handoffs, reassessing workforce needs as agents absorb diagnostic and remediation tasks, reviewing software, tooling, and managed-service contracts, removing redundant operational steps, and measuring cost of resolved operational outcomes. Bain also described Vivo in Brazil using a self-healing mechanism within Telefónica’s Autonomous Network Journey program for a virtualized standalone 5G core, including automatic detect-to-resolve actions and validation, and it said Telefónica reported a 30-minute reduction in mean time to resolution for targeted incidents. “Bain’s warning comes as operators accelerate plans for autonomous networks.”

In a separate element of the guidance, Bain said operators should evaluate end-to-end economics and govern agentic spending and behavior with limits on token, compute, and tool-call budgets, time limits for escalation to human operators, context-management rules, consolidation of overlapping checks, policy constraints for autonomous network changes, monitoring for model drift and spending anomalies, and business and financial ownership for each production agent and workflow.