It’s a wrap: News this week (March 2–6)

India’s early-March enterprise tech narrative was all about making AI dependable at scale. Across announcements between March 2 and March 6, the message was consistent: the next wave isn’t “more AI pilots”; it’s production systems, measured by user experience, optimized down to the chip layer, governed by data and semantics, and supported by serious infrastructure engineering built in India.

The digital workplace becomes the new control room

TCS and Zscaler pushed the AI conversation into the everyday reality of IT operations with the launch of the TCS Workspace Experience Studio, built with Zscaler Digital Experience (ZDX). The goal is straightforward: fuse zero trust security, digital experience monitoring, and AI analytics so IT teams can see—in real time—how apps, networks, and users are actually performing across distributed environments. The platform leans on Zscaler’s OneAPI to integrate services and adds observability plus automated remediation to detect and fix issues before they become outages. TCS also pointed to a deployment at a Scottish engineering company, projecting a 25–30% productivity lift based on internal estimates, and highlighted the shift from SLAs to “experience-level agreements” as CIOs start managing work as outcomes rather than uptime.

From pilot to production, now tuned for cost, performance, and governance

Infosys and Intel’s expanded partnership tackled the hardest part of enterprise AI: getting from demos to durable deployments. The collaboration pairs Infosys Topaz Fabric—positioned as a unified layer spanning infrastructure, data, models, applications, and workflows—with Intel’s compute portfolio including Xeon, Gaudi accelerators, and AI PCs. The emphasis is on jointly designing and optimizing workloads so enterprises can balance performance, security, and cost, especially in large-scale and regulated environments. Importantly, the partnership explicitly calls out AI agents operating within governance controls across data centers, cloud, and edge—signaling that “agentic” isn’t a future buzzword; it’s becoming part of the deployment blueprint.

Agentic AI for telcos: turning messy data into governed decisions

Tech Mahindra and Microsoft took the agent idea a step further with an ontology-driven agentic AI platform aimed at telecom operators and large enterprises. Built on Microsoft Fabric and Azure AI Foundry, the pitch is governance first: deploy and manage AI agents in complex environments while keeping outputs explainable and auditable. The platform is designed for a telecom reality shaped by mergers, legacy systems, and sprawling digital infrastructure—fragmentation that makes “one more AI tool” useless unless data is standardized and reusable. TechM says its approach converts enterprise metadata into structured data products, supports data-mesh adoption, and orchestrates multiple agents for use cases like churn prediction, fraud detection, revenue assurance, and network optimization. The semantic layer—mapping entities, relationships, and rules—aims to make AI decisions traceable and compliance-ready.

IBM’s Bengaluru bet: infrastructure engineering as India’s AI advantage

IBM launched its first Infrastructure Innovation Centre in India, the Sangam Infrastructure Innovation Centre, housed at a new India Systems Development Lab (ISDL) campus in Bengaluru. IBM describes it as a collaborative engineering hub where IBM specialists work with clients, ISVs, GSIs, GCCs, and partners to co-create secure, scalable, “responsible” enterprise AI systems—bringing hybrid cloud, infrastructure technologies, and AI platforms together. IBM also cited research from its Institute for Business Value: 58% of Indian organizations have increased infrastructure investments due to AI demand, with infrastructure budgets expected to grow by 19% in 2025, and 43% have established or are planning AI Centres of Excellence as they move from experimentation to scaled deployment.