Login
Register
Packet core operations are becoming more complex, but anomaly detection alone is not enough. This Catalyst uses multi-agent AI to turn data into insight, action and automation.
Detecting an issue is not enough. You have to understand it
Packet Core operations sit at the heart of mobile data, voice, IoT and enterprise services, making them one of the highest-risk domains in the network. As communications service providers (CSPs) move towards 5G and cloud-native operations, teams must assure services across Packet Core network functions, telco-cloud infrastructure, virtual machines, interfaces, topology and service context, while also interpreting telemetry, KPIs, alarms and logs from across that environment.
Many operators already use AIOps to detect anomalies and correlate events, but anomaly detection alone is not enough. Operations teams still need to understand what is happening, identify the likely root cause, assess potential service impact, decide the next best action and coordinate the right experts. That creates manual effort, inconsistent triage and continued reliance on scarce domain specialists during critical incidents.
The commercial impact is significant: slower response to service-affecting issues, greater operational risk, pressure on customer experience and less ability to scale operations as networks become more software-defined and distributed.
The Catalyst, Multi-agent intelligence for packet core operations, builds agentic intelligence on top of an existing production-grounded AIOps foundation for Packet Core operations. In that foundation, operational data from the mobile Packet Core is ingested, normalized and correlated with topology, infrastructure and service context, bringing together Packet Core telemetry, telco-cloud infrastructure data, network telemetry, topology models, metrics and events.
The multi-agent architecture then adds specialist roles. Network domain agents enrich detected anomalies with Packet Core expertise, drawing on logs, technical support files, configuration data, domain knowledge and root-cause hypotheses. Operational agents interpret the enriched situation, summarise the issue, analyse logs, recommend the best next action and support coordinated response through human-in-the-loop collaboration, including triggering automation where possible.
The result is a practical detect-enrich-reason-act-learn model for Packet Core operations that turns alarms, KPIs, logs, topology and historical ticket knowledge into an explainable operational summary, likely root cause, service impact context and recommended actions. The Catalyst aligns with key TM Forum assets, including Open Digital Architecture (ODA), TMF642 Alarm Management API, SID, Autonomous Networks guidance G1251 and IG1252, and IG1345 principles for responsible GenAI, explainability and human oversight.
Rather than stopping at anomaly detection, the solution helps CSPs move towards trusted, controlled autonomy in a high-risk network domain by combining operational data, domain intelligence and guided action in a standards-aligned operational flow.
For CSPs, the direct value is faster and more consistent handling of Packet Core incidents. Operations teams can move from manual interpretation of alarms, KPIs, logs and topology to an AI-supported workflow where the situation is summarised, likely root causes are identified, relevant context is surfaced and recommended actions are proposed.
The expected outcomes include reduced anomaly triage time, less manual analysis, better quality and consistency of root-cause insight, more effective guided resolution, faster cross-team collaboration and reduced dependency on expert escalation for recurring issues. This supports a safer path towards Level 4 autonomous operations.
For the wider industry, the Catalyst shows how CSPs can evolve existing AIOps and assurance capabilities into interoperable, multi-agent autonomous operations by building on trusted operational data, topology and service context rather than replacing what already works.
For society, the benefit is more resilient mobile connectivity. Packet Core reliability underpins mobile broadband, voice services, IoT connectivity and digital services used by consumers, enterprises and public-sector organisations, so faster and better-informed response helps reduce the risk, duration and impact of service disruption.
Marton Sabli, Head of Architecture & Roadmap, Telia Company adds, “This Catalyst is valuable because it demonstrates a practical path from today’s production AIOps capabilities to trusted autonomous operations. By combining Packet Core telemetry, topology and cloud infrastructure observability with cooperating domain and operational agents, operations teams gain a richer, service-aware view of incidents, including domain-level insight, likely root cause, recommended actions and collaboration support. For Telia, the direct benefit is faster and more consistent understanding of Packet Core issues, reduced manual triage and safer progress towards higher levels of autonomy in a domain where service impact can be significant. For the wider industry, the Catalyst shows how CSPs can evolve existing operations platforms with interoperable, human-supervised multi-agent intelligence.”
