Tech Mahindra, NVIDIA launch AI telco agent

Tech Mahindra has announced a collaboration with NVIDIA to develop an AI-powered telco network operations reasoning agent. The solution is intended to support communication service providers (CSPs) in progressing towards Level 4+ (L4+) autonomous networks by shifting traditional Network Operations Centres (NOCs) towards closed-loop operations.

Embedding AI into Network Operations

The offering is delivered through Tech Mahindra’s Orion platform and combines AI engineering with telecom domain expertise and human oversight. According to the company, the system is designed to operate without using customer or personally identifiable information (PII), with the aim of supporting data privacy and governance requirements.

Network Operations Centres have traditionally relied on rule-based processes and manual intervention, with engineers correlating alarms, logs and performance data across multiple systems. The companies state that embedding AI-driven reasoning into network operations can help automate alarm validation, root-cause analysis and incident resolution across Operations Support Systems (OSS) and Business Support Systems (BSS).

The solution enables CSPs to deploy a foundational large telco model and extend it with domain-specific reasoning agents integrated into existing data environments and workflows. Using NVIDIA AI Enterprise software, Tech Mahindra customised a reasoning model with synthetic and anonymised data through NVIDIA NeMo and deployed it as an NVIDIA NIM inference microservice. The company reports improved model accuracy compared with a non-fine-tuned version.

The collaboration reflects broader industry efforts to introduce AI-based automation into telecom network management through modular and scalable architectures.