Summary
Jeff Su walks through the four core capabilities of Gemini Spark, Google’s new proactive AI agent built into Google Workspace. The tutorial opens by drawing a sharp distinction between the standard Gemini chatbot—which responds when asked—and Gemini Spark, which acts autonomously: renaming uploaded files based on their content, routing and archiving emails on a set schedule, generating recurring reports from a template, and cross-referencing data across Gmail, Google Calendar, and Google Drive without manual prompting.
The video covers four capabilities in sequence: “Connect the Dots” (stitching information from multiple Google services into a single output), template-driven document generation, the Skills feature for building reusable automated playbooks, and personal preference memory. A standout example shows Spark drafting a calendar invite by identifying the correct contact in Gmail, reading the email thread for context, and automatically reusing the time slot from a prior meeting—all without the user providing a last name or looking anything up.
Setup instructions include enabling the Personal Intelligence memory toggle in Gemini settings, turning on Google Workspace connected apps, and creating a “Spark OS” folder in Google Drive to organize skills. Su also shares a downloadable inbox-zero skill file that can be imported directly into Gemini Spark, making the workflows immediately replicable for viewers.
📺 Source: Jeff Su · Published June 30, 2026
🏷️ Format: Tutorial Demo
