← Gemini for Workspace
Day 5 of 7✓ Sent
Google Agentspace — Building Custom Agents for Your Office
Agentspace is Google's no-code/low-code platform for building AI agents that connect to your Google Workspace data AND third-party tools. This is where you go beyond 'AI in a document' to 'AI doing work on my behalf'.
Key Points
- ✓What an agent does: Unlike Gemini in Gmail (which assists you), an agent can be triggered, complete multi-step tasks, query multiple data sources, and return a structured result — autonomously.
- ✓Built-in connectors: Agentspace has pre-built connectors to Gmail, Drive, Calendar, Google Meet, and a growing list of third-party tools (Salesforce, ServiceNow, Jira, Confluence, etc.).
- ✓Agent types you can build without code: Search agents (answer questions from your data), Task agents (take actions on your behalf), Workflow agents (run multi-step processes).
- ✓Real estate agent example 1 — Listing Agent: Trigger: new property added to spreadsheet. Agent: pulls property details, drafts listing copy in a Doc, creates a Slides presentation, schedules a review meeting in Calendar. One trigger, four outputs.
- ✓Real estate agent example 2 — Lead Qualifier: Trigger: new buyer enquiry email arrives. Agent: extracts buyer requirements, searches Drive for matching listings, drafts a personalised reply with 3 relevant properties attached.
- ✓Real estate agent example 3 — Compliance Checker: Agent reads a draft lease in Docs and flags any clauses that conflict with RTRA Act requirements (loaded from NotebookLM).
- ✓Low-code builder: Agentspace has a drag-and-drop builder. Vertex AI Agent Builder is the developer layer if you need custom code (Python, Node.js).