India’s global systems integrators build next wave of agents with NVIDIA AI

Tapping into NVIDIA AI Enterprise software and NVIDIA Nemotron models, India’s technology leaders are accelerating productivity and efficiency across industries — from call centres to telecommunications and healthcare. 

Infosys, Persistent, Tech Mahindra and Wipro are leading the way for business transformation, improving back-office productivity and customer services with integrated agentic AI platforms built with NVIDIA AI Enterprise.

- The Wipro WEGA Platform and NVIDIA AI Enterprise are jointly boosting efficiency for call centres, for instance. Wipro’s AI‑agent-assisted solution can handle peak enrollment cycles, when traditional contact centre business models choke, improving customer experience while containing the growing cost of service.

Already deployed for a major healthcare insurance provider, the system is already reshaping member experiences by enabling service representatives to handle more complex requests, accelerate resolution times, deliver more personalised support, and improve operational efficiencies.

AI agents help meet the expectations customers bring to their health plans: immediate access to accurate information, conversational self‑service, frictionless enrollment and consistent guidance across channels. Behind the scenes, payers face rising call volumes, fragmented data and heavy administrative workloads. AI agents bridge that gap by scaling instantly, operating around the clock and supporting human representatives with real‑time intelligence.

Today, 42% of inbound calls are handled by AI agents with near‑instant responsiveness across 900 concurrent calls and 164 requests per second — all with sub‑200‑millisecond latency.

Members benefit from natural, conversational self‑service. Human agents receive real‑time prompts and knowledge retrieval. A centralized data hub surfaces personalised insights, while automated digitisation removes manual work from downstream processes.

Using production grade, horizontally-scalable NVIDIA NIM microservices and NVIDIA NeMo Guardrails, part of NVIDIA AI Enterprise, the solution includes the performance, governance and safety required in regulated healthcare environments. Similar deployments under way in financial services.

- Tech Mahindra is accelerating the shift toward AI-assisted network operations with a new platform built in collaboration with NVIDIA. At the heart of the platform is a large telco model (LTM) that generates prioritised, data‑driven recommendations to help field technicians rank each fix by its historical success rate across the network. The result is faster, more accurate resolutions — often in a single visit — and a clear path toward level‑4‑plus operational maturity.

A large telecommunications services provider is adopting the same LTM foundation as part of its operations roadmap, targeting improvements in service‑layer issue resolution, customer experience and back‑office efficiency through higher‑quality tickets and fewer escalations.

The platform uses NVIDIA Nemotron, embedding models for semantic search across telemetry and a Nemotron re-ranking model to sharpen decision relevance. These models are deployed with NVIDIA NIM microservices for rapid, reliable accelerated AI inference. The NVIDIA NeMo Agent Toolkit orchestrates agent workflows across network domains, enabling agentic operations at scale.

- Infosys has developed a small language model (SLM) for coding using the NVIDIA NeMo framework that’s part of NVIDIA AI Enterprise. The SLM works within the Infosys Topaz Fabric to accelerate software delivery while remaining lightweight across on-premises enterprise data centres, cloud environments and even standard desktops.

The 2.5‑billion‑parameter SLM supports agent development, code generation, refactoring and end‑to‑end software‑engineering workflows. It is trained on high‑quality code, synthetic data, mathematical reasoning and natural language inputs — an approach that enables it to match frontier‑model performance on benchmarks such as MBPP, MBPP+ and BFCL.  

Safety and trust are prioritised. The model incorporates safety‑aligned training and responsible AI practices that reduce harmful outputs while preserving fluency. Its secure‑coding capabilities are validated through industry benchmarks such as the Stanford AIR‑Bench and Meta’s CyberSecEval, giving enterprises confidence to deploy it across code generation, debugging and multi‑agent development pipelines.

- Persistent Systems is working with NVIDIA to push early‑stage drug discovery into a new era of speed and scientific fidelity. The collaboration brings together Persistent’s deep life sciences engineering expertise with NVIDIA’s full‑stack accelerated computing platform, empowering researchers in everything from AI experimentation to production‑grade discovery workflows.

At the centre of the effort is Persistent’s new Generative Molecules and Virtual Screening (GenMoIVS) solution, built on the NVIDIA BioNeMo platform and the NeMo Agent Toolkit. GenMoIVS uses large, domain‑specific models to simulate molecular behaviour with high accuracy, generating and evaluating candidate compounds before they ever reach a wet lab. These agentic workflows continuously reason across virtual screening, prioritisation and experimental planning, helping teams de‑risk early discovery and shorten development cycles.

The platform runs on NVIDIA’s accelerated computing platform, including NVIDIA AI Enterprise software and NIM microservices, enabling high‑throughput simulation and real‑time scientific decision-making in regulated environments. By combining scalable infrastructure with production‑ready agentic AI, Persistent is giving life sciences organisations a faster, more cost‑effective way to explore the compound space and improve downstream success rates.

*A wet lab deals with hands-on experiments, as opposed to a dry lab which focuses on theoretical or digital work.