Erwan Menard on Scaling Enterprise AI with Google Cloud through Governance, Intentionality, and AI Agents | Cloud Wars Minute

Erwan Menard is the director of product management for Google Cloud’s Cloud AI division, where he helps lead innovation at the intersection of AI agents, enterprise systems, and business outcomes. In part two of our series, Google Cloud and the AI Revolution, Erwan joins Bob Evans to discuss how governance, intentionality, and rapid scaling are critical to AI agent success, share insights on Google Cloud’s Agentspace and Agent Builder tools, and explore how multi-agent collaboration is reshaping the future of enterprise technology.

Purpose Driven AI Innovation

The Big Themes:

• Intentionality Drives Impact: Menard advises organizations not to jump into AI agent development for novelty’s sake, but to begin with a clearly defined problem and desired business outcome. However, once value is proven, it's crucial to scale intentionally. He shares the example of a customer rolling out 40,000 licenses of Agentspace only after deeply considering what kind of first experience they wanted their employees to have.
• Organizational Culture Shapes AI Adoption: There's no universal model for who should “own” AI governance. It depends on the company’s culture. Some companies may create centralized AI governance teams; others may embed responsibilities within existing business units or IT teams. The key is cultural acknowledgment: governance must be understood as a shared responsibility, not just an operational afterthought.
• Anchor in Business Value: With so many tools, models, and frameworks emerging, it’s easy for companies to fall into what he calls “optionality evaluation.” That is, spending so much time chasing the latest innovations that they lose sight of why they started exploring AI in the first place. Instead, he urges leaders to ask: What are we trying to improve? Whether it’s speeding up contract workflows, freeing up data scientists from routine tasks, or enhancing customer service, the goal should be clear.

The Big Quote: "If you find yourself in a constant evaluation loop for the new shiny object, maybe it's worth taking a pause and saying, 'Why are we doing this again?'"