rewrite this title IT Leadership Interview: Why Enterprise AI Fails and How to Scale With HCLSoftware

Watch on Youtube.

In this session, UC Today’s Kieran Devlin sits down with Kalyan Kumar (KK), Chief Product Officer at HCL Software, to diagnose a critical issue facing the Global 2000: the inability to move AI from the lab to the real world. With so many firms stuck running endless experiments without delivering hard business outcomes, this conversation offers the architectural blueprint needed to break through the deadlock. KK shares why the secret to AI success isn’t actually about the AI itself—it’s about how you manage the data that feeds it.

Everyone is rushing to roll out AI, but few are seeing the productivity gains promised. Why? According to KK, it’s not an AI problem—it’s a data problem. The enterprise landscape is a “tangled web” of disparate applications, and without a data-first operating model, deploying autonomous agents often results in simply making bad decisions faster.

In this deep dive, we explore why modernization doesn’t mean ripping out “classic” systems like mainframes, but rather building an orchestration layer that connects them to new intelligence. KK explains why the future isn’t just about picking an LLM, but about mastering metadata and preparing for a multi-agent world where governance is non-negotiable.

Key discussion points:

The Data-First Imperative: Why you must untether data from applications and master metadata before AI can succeed—treating your enterprise like a well-organized library rather than a chaotic storage room.

Solving the Integration Paradox: How to bridge modern AI agents with “classic” core systems (ERPs, Mainframes) using universal orchestration rather than forcing a total rip-and-replace modernization.

Governance in a Multi-Agent World: Preparing for the rise of Agent-to-Agent (A2A) communication and the Model Context Protocol (MCP) to prevent autonomous agents from creating conflict.