During Google Cloud AI Live in Amsterdam we spoke with Simon Lebrun, product manager Cloud AI at Google Cloud, for a deep dive into Gemini Enterprise. Simon explains how Google’s AI strategy connects enterprise applications, APIs, and emerging protocols like MCP and A2A to bring meaningful AI to every employee, with or without a technical background.
The conversation covers why API management and iPaaS are fundamentally different disciplines, how Apigee acts as a governance and distribution layer for MCP servers, and how Gemini Enterprise bridges Microsoft 365, Salesforce, Workday, Slack, and virtually any enterprise system through a single pane of glass. Lebrun also addresses the tension between open AI ecosystems and vendors building new walled gardens, and explains Google’s approach to supporting both.
Finally, Lebrun introduces Spark, Google’s latest AI agent generation that acts as a personal 24/7 assistant, understanding user intent, proposing plans, and orchestrating specialized agents automatically. He also tackles the challenge of data volume: how search, agent grounding, and orchestration work together to keep AI responses accurate without overloading models with too many tools or sources.
🔑 Key takeaways:
• APIs are the non-negotiable foundation of enterprise AI: internal and external
• API management (Apigee) and iPaaS serve different purposes and should remain decoupled
• Gemini Enterprise app connects Microsoft 365, Salesforce, Workday, Slack, and any MCP server
• MCP is for tools, A2A is for agent-to-agent interaction, governed by the same principles
• Apigee distributes MCP servers in optimized subsets (e.g., 10 tools for HR, 20 for supply chain)
• Too many tools slow agents down: governance and distribution are critical
• Spark orchestrates specialized agents automatically based on user intent
• The industry is moving from conversational chatbots to structured, deterministic AI agents
⏱ Chapters:
0:17 Introduction: Gemini Enterprise and Google Cloud AI
0:57 Why APIs are the backbone of enterprise AI
1:51 API management vs. iPaaS: Apigee explained
3:02 Gemini Enterprise app in the Microsoft ecosystem
4:17 Open platforms vs. walled gardens: MCP and A2A
8:18 Governance and observability for MCP, APIs, and A2A
10:54 Managing data sources: search, NotebookLM, and agent grounding
13:59 Spark: Google’s personal AI assistant for enterprise
15:03 From chatbot to deterministic AI agents in enterprise
Keywords: Gemini Enterprise, Google Cloud AI, Apigee, MCP, A2A, API management, enterprise AI, AI agents, Spark, agent orchestration, iPaaS, enterprise integration, Google Cloud, AI governance