Dr. Dale Skeen, CTO & Co-Founder , Vitria Technology
AI is powering the journey toward Self-Healing Networks – but current AI agents and approaches are not yet accurate enough to achieve a high-level of network autonomy. Vitria is pioneering a new approach within our VIA AIOps product line, where AI is augmented with Knowledge to drive superior results and enable new levels of Autonomy.
At the heart of this new approach lies “structured knowledge”, often represented as a “Knowledge Graph,” a resource that is both machine and human readable. The Knowledge Graph is a sharable accumulation of the knowledge critical to detecting, diagnosing, and remediating outages and incidents occurring in large telecom networks.
VIA AIOps constructs Knowledge Graphs from multiple sources, including real-time monitoring data from IT infrastructure, networks, cloud environments, and applications, and from contextual data stored in CMDBs and Network Inventory Systems. Knowledge is further augmented with historical data such as incidents, resolutions, and automated actions, ensuring that decision-making accuracy is continuously refined and improved.
The Knowledge Graph serves a shared asset for various AI Agents and algorithms, access either directly or through connectors and “transformers” that provide Knowledge in AI-consumable forms. This design allows all AI Agents – from traditional AI/ML models to advanced LLMs – to utilize the Knowledge Graph.
This “AI + Knowledge” approach has enabled self-healing operations for Vitria’s customers, including dynamic rerouting of traffic during fiber cuts, link failures, and congestion; rollback to a last known good state for erroneous or impactful configuration changes; and dynamic resource allocation to manage traffic spikes.
The key capabilities and benefits from Vitria’s AI + Knowledge approach include:
- Elevated AI Accuracy and Contextual Relevance. By tapping into the Knowledge Graph, AIs significantly improve prediction quality and overall effectiveness.
- Collective Intelligence Through Shared Knowledge. Once an AI agent learns a concept, that knowledge is shared with other agents, creating multiplicative gains and virtuous cycles of reinforced learning.
- Synergy Between Human and Machine. The Knowledge Graph is openly built and shared among AIs and human operators, fostering continuous, two-way feedback that combines expert insight with AI-driven validation and improvement.
- Autonomous Knowledge Acquisition. Specialized AI Agents build, maintain, and validate the Knowledge Graph, accelerating independent knowledge growth and minimizing manual overhead.
- Service-Based, Open, and Federated. VIA AIOps Knowledge Plane supports multi-agent architectures, and open Knowledge sharing with other functions—such as Capacity Planning and Service Provisioning.
- Accelerated Resolution and Continuous Improvement. A knowledge-based approach reduces mean time to detect and resolve issues, and enables persistent learning, ensuring the Knowledge Graph continuously evolves over time.
Knowledge Graphs significantly improve Generative AI outputs by introducing structured, verified information, thereby making generative AI more accurate, explainable, and reliable. The graph’s explicit relationships and semantics help the LLM follow a more rigorous “chain of reasoning” when answering queries, rather than relying solely on pattern-based associations learned from text. Because Knowledge Graphs are updated and verified independently, they serve as a robust guardrail, ensuring model responses remain consistent with established facts.
Why is this important? As the industry moves towards Network as a Service, driven by technologies like 5G and network slicing, improvements in self-healing and autonomous networking are essential for overcoming the increasing complexity and costs of modern networks. VIA AIOps with Knowledge provides the innovation that enables a new level of self-healing operations for Vitria’s customers, including:
- dynamic rerouting of traffic during fiber cuts, link failures, and congestion;
- rollback to a last known good state for erroneous or impactful configuration changes;
- optimizing bandwidth allocation in real-time;
- and dynamic resource allocation to manage traffic spikes.
By infusing knowledge-based intelligence, context-aware decision-making, and autonomous healing into AIOps, VIA AIOps with Knowledge is transforming AI-driven operations.
Vitria Technology is a Gold Sponsor at FutureNet World 2025. Meet Dr. Dale Skeen and the team on 7-8 May, London.
