Omdia, April 2026. Author: Ruth Brown, Principal Analyst, Mobile Networks. Commissioned by Vitria Technology. The analysis and conclusions are Omdia’s.
What the paper argues
Traditional systems are built on data persistence. They record what happened. A knowledge-based system instead treats the relationships between things as primary data in their own right, which turns siloed records into something an agent can reason over. Omdia’s case is that this is the difference between a decision made on statistical inference and one made with an understanding of dependency and impact.
Why it matters for multi-agent systems
Decision-making in operations is moving away from monolithic scripts toward distributed agents that act independently and collaborate on larger problems. That only works if the agents share an accurate picture. Without one, they duplicate work, conflict with each other, and cannot coordinate. The paper sets out why a knowledge model that continuously learns and updates is the precondition for agents working together rather than in parallel.
What an open knowledge plane does
Omdia describes the process in three stages: acquiring heterogeneous source data, both structured and unstructured; organizing it against standardized telecom ontologies such as TM Forum and ITU-T so that entities and relationships are consistently defined; and exposing it through shared services so that multi-vendor agents and workflows all read from the same understanding. The paper compares this directly against legacy relational approaches across structure, context, scalability, explainability and governance.
A worked example
The paper follows a performance degradation in a 5G network from detection through to resolution: a monitoring agent identifies anomalies and uses the knowledge graph to establish blast radius; a diagnostic agent examines dependencies across cell congestion, transport and compute; the system reasons over what changed and checks policy constraints before acting; and the graph is updated with the outcome so the next similar incident resolves faster.
“Knowledge will become a foundational capability for multi-agent automation.”
Who should read it
Network and IT operations leaders evaluating agentic AI, and anyone being asked to distinguish a genuine multi-agent system from a rebadged automation platform.
This Omdia white paper was commissioned by Vitria Technology. © 2026 TechTarget, Inc. All rights reserved. The analysis and conclusions are Omdia’s.
