CASE STUDY 02 / GOOGLE SHEETS

GEMINI
INTEGRATION.

Defining how Generative AI lives inside the spreadsheet: From the Gemini function to AI-powered Tables.

SCOPE
3 AREAS

FUNCTION, TABLES & STRATEGY

ROLE
STRATEGY

DEFINED CAPABILITIES & TRAINING DATA

VALIDATION
TESTING

MULTIPLE ITERATIVE ROUNDS

THE CHALLENGE

Structuring Intelligence

Integrating GenAI into a structured tool like Sheets required more than just a chatbot sidebar. The challenge was to make AI native to the grid. I led the foundational research for the Gemini-powered AI function, helping shape its initial capabilities and defining the training examples that taught the model how to understand spreadsheet data.

RESEARCH METHODS
  • MODEL TRAINING EVAL SURVEYS
    USE CASES, NEEDS & SAMPLE PROMPTS
  • PROMPT ENGINEERING MODERATED
    USABILITY TESTING
  • INTERFACE USABILITY HYBRID
    MODERATED & UNMODERATED
STRATEGIC PILLARS

THE AI FUNCTION

Conducted foundational work to define the capabilities and interface for the Gemini-powered function, enabling natural language processing directly in cells.

AI TABLES

Leveraged foundational data to shape AI Columns and AI-embedded templates, automating the creation of structured tables.

[CONFIDENTIAL]

Spearheading research for a future strategic AI capability (details under NDA).

FRANCISCO
VELASQUEZ

BUILD: 2026.0.1

© 2026 ALL RIGHTS RESERVED.

LOCATION DATA

NEW YORK, NY

40.7295° N

74.0064° W

CONNECT