Enterprise leaders on agentic AI at Google Cloud roundtable: Trust, guardrails, and bold bets

Enterprises are entering a new phase of their AI journey. The early rush of pilots, chatbots, and productivity hacks has given way to a more pressing question: how can businesses move beyond automation and unlock true enterprise value with AI-driven agents? For leaders grappling with this shift, the opportunities are enormous, but so are the challenges of governance, trust, scalability, and cultural readiness.

These dilemmas shaped the agenda of Beyond Automation: Unlocking Enterprise Value with Agentic AI, a closed-door roundtable hosted by Google Cloud, in association with YourStory, on the sidelines of DevSparks 2025. Moderated by Ipsita Basu, Senior Director at YourStory, the discussion brought together leaders from across banking, healthcare, manufacturing, and technology to share where agentic AI stands today, what is working, and what remains stubbornly out of reach.

The panel featured Guru Bala, Customer Engineering Manager at Google Cloud; Kalyan C Varanasi, AVP Cloud and IT at Infosys; Rama Kumar Nadipalli, Account CTO and Head of Architecture at HCL; Hari Krishna Reddy Abba, Director of Product Management at Pine Labs; Abhinav Pachauri, Vice President of Product at Kotak Mahindra Bank; Sampath A, Senior Vice President of Technology at Axis Bank; Gaurav Kataria, Vice President – Digital and CDIO at ITC; Kiran Kumar Sanapathi, AVP Data Analysis at IndusInd Bank; Madhu Madhanan, VP and CTO of Platform Engineering at CitiusTech; Rajesh Unde, CTO at PL Capital; Anand Reddy Kukudala, Co-founder and CTO at Tericsoft; and Pankaj Khare, CIO at IndusInd Bank.

The value question

A thread running through the discussion was the elusive definition of “value”. Panelists stressed that small efficiency gains—say, shaving 20% off a process—are not enough to justify the scale of AI investment. For agentic AI to matter, it must move the needle on growth, cost efficiency, compliance, or customer experience in a way that is visible and measurable.

Several participants admitted that many AI pilots fail because they are bolted onto outdated, human-centric workflows. Simply digitizing an old process won’t deliver transformation. The consensus: enterprises need to reimagine workflows with agents at the core, not as add-ons, to unlock genuine business value.

Banking leaders pointed to the fact that customers still seek the reassurance of human interaction, especially when it comes to sensitive decisions like lending or fraud checks. The group agreed that AI agents should act as enablers, empowering employees to act as experts while governance, guardrails, and clear handoff points remain non-negotiable.

Scaling beyond pilots

Across industries, the panelists highlighted a shared “pilot problem”. Most companies are running multiple experiments, but few have managed to scale them. Incremental improvements may be useful, but they rarely create enterprise-wide momentum. The roundtable agreed that to demonstrate real value, organizations must take on bold, high-impact use cases even if that means higher risk.

Participants aThe bigger picture

lso acknowledged workforce anxieties, particularly in large engineering organizations. With thousands of developers and operations staff in play, the fear of displacement is real. The discussion emphasized that agentic AI should be framed as an enabler capable of boosting productivity by 30–40% rather than a threat. Reskilling and cultural change will be critical to that narrative.

Finally, the economics of adoption dominated the later part of the session. While boards and CXOs demand immediate returns, the consensus was that agentic AI requires patience. Iteration, refinement, and alignment with business logic are essential before enterprises can see compounding benefits at scale.

As Google Cloud’s Guru Bala and the other panelists underscored, moving beyond automation is not about building in-house large language models or chasing every shiny tool. It’s about rethinking processes, embedding trust, and preparing organizations, culturally and structurally, for an agent-driven future.

Enterprises that treat agentic AI as business transformation, rather than technology experimentation, will be the ones to unlock true enterprise value.

.