Every organization is at a different stage in the AI journey, whether just getting started or actively scaling impact. Mahalakshmi Nageswaran opened the conversation by framing the central challenge: How to move from AI ambition to tangible outcomes. From there, panelists shared real-world stories of how AI is being operationalized across industries to improve productivity, boost efficiency, and embed intelligence into everyday decision-making.

Swetha Sunkara from Dell Technologies picked up on that theme, noting that every company sits somewhere along the AI maturity curve. What separates leaders from laggards is clarity. Clarity in use cases, alignment across teams, and discipline in execution. She emphasized the importance of defining a North Star and grounding it in what is feasible versus what is aspirational. She shared that her team’s AI adoption journey began with exploring foundational building blocks for AI applications and scaled across the broader product management team, with domain experts embedding AI components, ranging from LLMs to agent-based automations, into areas such as onboarding, infrastructure monitoring, and incident management. 

Vishal Mahna of AVEVA brought the industrial perspective, where physical systems and digital twins converge. It is core to solving practical challenges like energy efficiency, machine health, and operator safety. He shared how AI-powered predictive analytics anticipates faults and unplanned downtime, cutting costs and operational risk by monitoring real-time lubrication, thermal, and process data. 

Lee Davidson shared lessons from AssetMark’s financial services transformation. One of the biggest hurdles is getting people to trust the AI’s output. He explained that it is not enough to show results. Teams have to show their work. Data lineage and explainability are crucial for building confidence, particularly in a regulated environment. That is why his team leans into transparency and ongoing validation as part of their operational AI strategy.

Sharleen Sy of Adobe offered a marketer’s lens. She spoke about how her team uses AI for experimentation and campaign optimization, but warned against blindly chasing automation. Just because you can automate does not mean you should. One challenge she sees is around “AI slop”, generic and repetitive outputs that erode brand value. Her team works to ensure that AI-powered personalization still reflects Adobe’s creative standards.

Mahalakshmi returned to the theme of cultural readiness, highlighting a shared concern about how organizations are rushing to implement AI but often lack the internal structure, skills, and change management needed to support it. Panelists agreed that communication and cross-functional alignment are just as critical as the technology itself.

A key theme across the panel was the need for guardrails. Whether in the form of governance, ethics, or documentation, successful AI initiatives require intentional boundaries. As Swetha noted, AI will fail without a framework. That includes deciding what not to automate, identifying failure points early, and keeping humans in the loop for high-stakes decisions.

Vishal shared that AVEVA has implemented a formalized set of controls known as “fail-safe mode” to ensure that AI systems operate within defined parameters in complex industrial environments. It is not just about launching a model. It is about making sure it behaves predictably in the wild.

Lee emphasized the importance of ownership. At AssetMark, each model has a “steward” responsible for performance, maintenance, and escalation. AI is not a set-it-and-forget-it system. It is a living product that evolves.

Sharleen closed the session with a reminder that successful adoption is an iterative process. You will not get it perfect out of the gate, but if you are learning, you are progressing.

The panel echoed a larger trend in enterprise AI, which is a shift away from hype and toward operational excellence. Proving AI can work is not enough. Success is about proving it can scale responsibly, reliably, and with real-world outcomes.