Everyone wants to deploy AI. Almost nobody wants to do the work that makes AI actually function.
Manjula Mahajan has spent 25 years doing that work — first as an integration architect making enterprise systems talk to each other, then leading data and AI transformations across Symantec, Brocade, Broadcom, NetApp, and Model N. Today she is VP of Enterprise Technology, Data & AI at Roller, a global venue management platform operating across 30 countries.
Her point of view is direct: AI doesn't fix broken systems. It exposes them. And the organizations that learn that lesson after the fact pay a much higher price than the ones that build the foundation first.
In this episode, we cover how a GenAI implementation generated over 100,000 hours of productivity gains — not by building AI features first, but by fixing the knowledge base underneath them. We get into her value-vs-risk matrix for prioritizing AI use cases, why she reframes governance as the guardrail that lets you move fast rather than the rule that slows you down, and what 11 years at one company taught her about watching architectural decisions actually play out over time.
The throughline across all of it: find the business problem. Everything else follows.
