Company Announcements

BARCELONA, Spain, March 3, 2026 /PRNewswire/ -- Tech Mahindra (NSE: TECHM), a leading global provider of technology consulting and digital solutions to enterprises across industries, announced a collaboration with NVIDIA to introduce an industry-first AI-powered Telco Network Operations Reasoning Agent. The solution is designed to help Communication Service Providers (CSPs) accelerate their journey toward Level 4+ (L4+) autonomous networks by transforming traditional Network Operations Centers (NOCs) into intelligent, closed-loop operations.

The solution is delivered through Tech Mahindra's Orion platform, combining AI engineering, domain expertise, and human-in-the-loop design. Additionally, it ensures enterprise-grade trust without using customer or Personally Identifiable Information (PII) allowing CSPs to safely operationalize artificial intelligence (AI) while maintaining data privacy and governance. The collaboration highlights Tech Mahindra's capabilities in helping operators advance toward L4+ autonomy through scalable, production-ready AI frameworks.

Amol Phadke, Chief Transformation Officer, Tech Mahindra, said, "Network operations centers still rely on rule-based, open-loop workflows with significant manual intervention. Engineers continue to spend considerable time correlating alarms, logs, and performance data across systems, impacting resolution times and operational efficiency. Our collaboration with NVIDIA addresses this challenge by embedding AI-driven reasoning into the core of network operations. By enabling intelligent, closed-loop execution, we are helping Communication Service Providers (CSPs) move decisively toward simplified operations aligned with L4+ autonomous network ambitions."

The AI reasoning agent within the solution enables CSPs to move beyond conventional automation by embedding contextual intelligence directly into network operations. It supports autonomous alarm validation, root-cause analysis, and resolution across Operations Support Systems (OSS) and Business Support Systems (BSS) environments, enabling faster incident handling, greater operational consistency, and a better customer experience. A key differentiator of the solution is its extensibility, as once CSPs deploy a foundational Large Telco Model, they can build and scale domain-specific reasoning agents that integrate with their existing data lakes, tools, and workflows. Leveraging NVIDIA AI Enterprise software, Tech Mahindra customized a reasoning model with synthetic and anonymized data using NVIDIA NeMo and deployed it as NVIDIA NIM inference microservice. The model provided 2-3x improvement in accuracy compared with a non-fined tuned model.

Chris Penrose, Vice President, Global Business Development – Telecom, NVIDIA, said, "Network operations demand rapid decision-making across complex, real-time environments. By combining NVIDIA's AI software stack with Tech Mahindra's deep telecom expertise, this collaboration enables CSPs to deploy reasoning-based AI systems that can act, adapt, and learn within live NOC environments. Together, we are laying the groundwork for a new era of autonomous, resilient, and intelligent networks."

The transition to autonomous networks is an iterative and ecosystem-driven journey, as modular architecture and reusable pipelines enable CSPs to adopt AI reasoning incrementally and scale it across their operations. The collaboration reinforces Tech Mahindra's commitment to driving AI innovation in telecom and redefining network operations through intelligent automation, deep learning, and multimodal AI models.

For more information on how TechM can partner with you to meet your Scale at Speed™ imperatives, please visit https://www.techmahindra.com

Media contact:
Abhilasha.Gupta@TechMahindra.com
media.relations@techmahindra.com

Logo: https://mma.prnewswire.com/media/2804444/Tech_Mahindra_New_Logo.jpg

 

View original content:https://www.prnewswire.co.uk/news-releases/tech-mahindra-advances-ai-driven-autonomous-network-operations-for-csps-globally-with-nvidia-302702436.html