Mumbai’s chapter of Dell Tech Forum 2025 focused on the tough questions that define enterprise AI. How do large and regulated companies move from proofs of concept to full-scale production? How can models stay explainable, compliant, and cost-effective? And how can infrastructure be built to handle the next generation of chips, tools, and regulations without causing disruption? To explore these themes, Forbes India partnered with Dell Technologies and hosted a session on enterprise AI adoption, moderated by Mridu Bhandari, anchor of Forbes India, with insights from Abhishek Sukhwal, Head of Infrastructure (Cloud) at Mahindra Group; Sriram Sridharan, Associate Partner at Lakshmikumaran & Sridharan Attorneys; Shankar Subramanian, General Manager – AI Sales, Asia Pacific & Japan at Dell Technologies; and Neelakantan Venkataraman, Vice President and Global Head of Cloud at Tata Communications.
Advertisement
The conversation settles quickly on first principles. Enterprise AI earns its keep through reliability, clarity, and scale. Shankar Subramanian offers a framework that reads like an operating manual. “Data quality, skills, security, the right partners, and robust infrastructure are the five pillars,” he says. “Our AI Factory is built on these pillars, so deployment becomes easier, scale is possible, and the platform stays sustainable.” He notes that Dell engineers have logged more than 500,000 hours validating designs and produced over 200 reference architectures tailored to Indian conditions. The signal is steady: plan for production from day one.
Within Mahindra, this plan takes the form of a cross-group platform. The company’s spread across automobiles, finance, and hospitality turns every interaction into operational intelligence. “Earlier, we could only sample audio calls at customer centers. Now with multimodal models, we can transcribe and analyze every single conversation, understand what is happening with vehicles, dealers, and services across the country, and act on it in real time,” says Abhishek Sukhwal. He describes voice agents that log issues through connected cars and flow straight into workshops, and a data program that pulls legacy systems into a governed fabric. “AI is not a vertical,” he adds. “It cuts across manufacturing, supply chains, customers, and employees.” The emphasis is on compounding value through shared foundations.
Neelakantan Venkataraman from Tata Communications emphasizes that the core of AI is how data moves. He points out that latency, compliance, and sovereignty are practical concerns, not just abstract risks. “We built GPU infrastructure within the country so enterprises can experiment, fine-tune, and productionize models without any metadata leaving Indian soil,” he explained. His team has developed AI Studio, a platform enabling companies to bring their data, govern it, and train models in secure environments. Their approach is practical: cloud services optimized for speed, safety measures built for trust, and flexible architectures prepared for future hardware upgrades without issues.
Advertisement
Law provides a unique perspective and a shared desire for accuracy. “GenAI will revolutionize law much like computers and search engines did in the early 2000s,” states Sriram Sridharan. His firm is custom-training a model on forty years’ worth of its own documents to improve recall, drafting, and review processes. Unlike generic systems, which lack a firm’s history, a tailored model can incorporate it. Sridharan highlights the influence of the current regulatory environment on technological choices. With new data protection laws and expanding digital legislation on the horizon, firms need to demonstrate data sources, usage by models, and explain decision-making. Here, innovation and accountability are linked.
Read More
Attention shifts to the next frontier: agentic systems that plan, decide, and act within set boundaries. Subramanian outlines the operating rhythm. Sales teams enter conversations with the right context. Support teams ensure follow-ups are completed on time. Engineers troubleshoot using guidance from the organization’s memory. “The future of work will be defined by the relationship between humans and agents,” he says. Dell is adjusting accelerators with partners and aligning software stacks so models that generate can sit alongside those that reason and call tools. The goal is seamless integration with enterprise workflows.
A consistent theme throughout the discussion is a strong emphasis on governance. Provenance, lineage, auditability, and cost discipline are considered core design elements rather than afterthoughts. Teams seek platforms that efficiently accommodate new chips and frameworks without waste and architectures that clearly define data boundaries. The culture surrounding AI evolves when these expectations are integrated from the beginning.
Advertisement
The pages slugged ‘Brand Connect’ are equivalent to advertisements and are not written and produced by Forbes India journalists.
The pages slugged ‘Brand Connect’ are equivalent to advertisements and are not written and produced by Forbes India journalists,The pages slugged ‘Brand Connect’ are equivalent to advertisements and a
