Ericsson has restructured its OSS and BSS customer support operations to accommodate agentic AI capabilities, signalling that enterprise telecom vendors are moving beyond chatbot-assisted triage into autonomous agent deployment. This shift reflects a broader industry recognition that traditional support architectures—built around human-first workflows with AI as a supplementary layer—are becoming obsolete. The update positions Ericsson's support function to handle the operational complexity that emerges when autonomous agents begin making decisions, escalating issues, and managing customer interactions without human intervention at each step. For CX teams already managing Zendesk or Salesforce Service Cloud instances, this raises an immediate question: are your current permission models, audit trails, and escalation protocols robust enough to handle agents that operate independently between human touchpoints?
The implications extend beyond Ericsson's own operations. This move validates that agentic AI is no longer a proof-of-concept concern for support leaders—it's an infrastructure requirement. Teams will need to reconsider how they architect knowledge bases, define agent boundaries, and measure success metrics when the unit of work shifts from "tickets resolved by humans" to "interactions managed by autonomous systems." The challenge isn't deploying the technology; it's redesigning support operations to accommodate agents that can handle multi-step problem resolution, context retention across sessions, and dynamic decision-making. For mid-market support teams currently optimising around first-contact resolution rates and handle time, agentic AI introduces a fundamentally different performance model where the relevant metric becomes "resolution autonomy"—the percentage of customer issues resolved without human intervention. This demands not just new tools but new operational governance, training frameworks, and quality assurance approaches that most teams haven't yet developed.
Ericsson updates OSS and BSS customer support in light of agentic AI revolution Telecompaper