Infosys announced the expansion of its collaboration with NVIDIA for AIpowered, customer-centric solutions to drive innovation and operational excellence for telcos.

Leveraging Infosys Topaz, an AI-first set of services, solutions, and platforms using generative AI technologies, the collaboration will help telcos enhance their customer experiences, streamline network operations, and accelerate service delivery.

Infosys developed three generative AI solutions, all of which are powered by Infosys Topaz, using NVIDIA NIM inference microservices, NVIDIA NeMo Retriever embedding models, and NeMo Guardrails to customize and deploy generative AI telco domain-specific LLM models. Infosys Topaz also uses NVIDIA Riva, for building real-time conversational AI pipelines that allow for real-time transcription and translations for call center agents.

- TOSCA Network Service Design leverages NVIDIA generative AI to create and customize TOSCA (Topology and Orchestration Specification for Cloud Applications) templates. This solution streamlines the network service design process. Using NVIDIA NIM and NeMo Retriever microservices resulted in a 28% reduction in latency and 15% absolute improvement in accuracy.

- An NVIDIA generative AI-powered smart networking operating center allows network engineers and network operations personnel to troubleshoot issues faster and accurately. With NVIDIA NIM and NeMo Retriever, this solution reduces mean time to resolve network issues and improve customer experience, resulting in 61% lower latency and a 22% absolute improvement in accuracy.

- Infosys Cortex – A generative AI-driven contact center transformation platform. Leveraging Infosys Topaz and Infosys' cloud suite of solutions, Infosys Cobalt, Infosys Cortex creates realistic customer simulations so that trainee agents can practice their skills in a completely immersive environment.

These solutions can bring tremendous benefits for customers in terms of operational efficiencies and productivity improvement and enhance the overall experience of contact center agent training by optimizing the end-to-end generative AI pipeline and inference costs.