Google Cloud VP reveals how to build an AI-native marketing team at scale

The Gist

• Sarah Kennedy Ellis says workflow friction, not AI quality, is the biggest adoption blocker
• Top 20% of AI adopters at Google are also the most trained, creating a productivity gap
• Training must fit into 5-minute weekly chunks to drive real adoption

• The greatest friction in a workflow is the biggest inhibitor to AI adoption, not model quality.
• Top 20% of AI adopters within Google Cloud's marketing team are the most productive due to deliberate skill-building.
• AI Boost Bites: 5-to-7-minute training videos drive adoption by fitting into the limited time employees have for learning.
• High-volume tasks with limited human judgment required are the best use cases for AI agents.
• Marketing's role is shifting from defending functional boundaries to leading technology that produces more effective sellers.
• Governance of AI agents should focus on shared infrastructure rather than stifling innovation.

• Implement bite-sized training (5-7 minutes) tailored to your team's workflow gaps to drive AI adoption.
• Focus AI deployment on high-volume, low-judgment tasks where quality can scale with output.
• Hire for curiosity and hands-on AI experience (e.g., 'what you built') rather than traditional management skills.
• Plan for agent governance early by building shared infrastructure to enable cross-team reuse of successful agents.

• 70% (Reduction in production time for creative assets during the Gemini in Chrome launch.)
• 5 minutes (Time employees typically have per week for learning new AI tools.)
• 18 months (Duration Google allowed unrestricted experimentation with AI agents before implementing governance.)