A multi-agent ecosystem built for telecom scale
Nokia is introducing six specialized agents, each designed to manage specific operational tasks independently or work together to solve complex network problems, including:
Router agent: Acts as the central orchestration layer by interpreting user intent and managing communication between other agents while ensuring compliance with operational guardrails.Event triage agent: Analyzes ongoing alarms and compares them against historical patterns to identify root causes and assess operational impact.KPI selector agent: Provides domain-expert interpretation of complex network performance metrics, definitions, and measurement units to aid reasoning.Anomaly reasoner agent: Investigates unusual network behavior to determine if a deviation is a genuine issue or a false alarm.Action reasoner agent: Matches active events against automation catalogs to recommend specific remediation steps.Dashboard agent: Enables teams to quickly generate visual analytics and tracking screens using natural language prompts.
Nokia developed these capabilities using Google Cloud’s Agent Development Kit (ADK) on the Gemini Enterprise Agent Platform. This framework allowed Nokia to quickly develop smart conversational and routing paths that leverage Gemini’s multimodal reasoning. Instead of requiring complex managed services, the entire multi-agent framework runs seamlessly on standard Google Cloud compute and storage. By deploying the agents directly via standard tools like Kubernetes and Google Cloud Storage, Nokia ensures the solution remains cost-optimized, flexible, and fully compatible with existing customer environments, including Google Cloud tenants.
Rather than removing operators from the loop, the framework introduces what Nokia calls “glass box autonomy.” The action reasoner agent serves as an advisory layer, presenting confidence-based recommendations to human engineers, who retain final approval over critical control points before fixes are automatically executed and logged. For low-risk, policy-approved scenarios, the same architecture can also support completely closed-loop automation.
Business benefit for operators
By moving from passive monitoring to a proactive, agent-based architecture, telecom providers can save costs and improve reliability. These specialized agents can deliver immediate value, including:
Rapid resolution: Agents can automate troubleshooting, cutting network problem-solving times by 50% to 80%. Complex issues that historically took hours to isolate—such as voice degradation or software errors—can now be flagged and resolved in minutes, drastically reducing service downtime.Enhanced accuracy: Agents can filter out data fluctuations to focus on genuine problems, reducing false alarms and unnecessary escalations.Accessible analytics: Staff can use simple language to instantly generate dashboards and performance reports.Cost efficiency: Agents efficiently run on top of a telecom operator’s existing cloud setups, avoiding the need for expensive custom software overlays.
‘;