Building My Own Personal AI Assistant: A Chronicle, Part 2

2026-04-16 · Source: Towards Data Science · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, medium

Summary

The second part of the "Fernão" personal AI agent development focuses on significant architectural and feature enhancements. The calendar fetching mechanism was refactored from inefficient ICS parsing to a Google Calendar API integration, reducing schedule generation time from nearly five minutes to about twenty seconds by enabling native event filtering. Beyond backend improvements, new user-facing features include the ability to mark tasks as completed, which automatically syncs with Microsoft To-Do. A new "Task Breaker" module was introduced, allowing users to input large, generic tasks with context, which Fernão then decomposes into smaller, actionable 20-minute subtasks with suggested due dates, saving them back to Microsoft To-Do. The article also previews upcoming modules like a Dividend Analyzer and Writing Assistant.

Key takeaway

For AI Engineers building personal agents or productivity tools, consider migrating from generic data formats like ICS to native APIs such as Google Calendar API to drastically improve performance and enable advanced filtering. Implementing an AI-powered task decomposition module, like the "Task Breaker," can significantly enhance user productivity by transforming large projects into actionable, time-boxed subtasks, directly integrating with existing task management systems like Microsoft To-Do.

Key insights

API-driven data access and AI-powered task decomposition significantly enhance personal agent efficiency and utility.

Principles

• Prioritize native APIs over generic formats for efficiency.
• Modular design enables flexible integration and feature expansion.

Method

The Task Breaker workflow involves providing a large task and context to an AI agent, which then decomposes it into 20-minute subtasks with due dates, saving them to a task management system.

In practice

• Integrate Google Calendar API for efficient event retrieval.
• Use AI for breaking down large projects into manageable subtasks.
• Sync task completion across different productivity apps.

Topics

• Personal AI Assistant
• Google Calendar API
• Task Management
• Microsoft To-Do
• Task Decomposition

Best for: AI Engineer, Software Engineer, Prompt Engineer

Related on AIssential

• The End of Apps — Kitze, Sizzy.co — The pursuit of a perfect "life OS" reveals persistent friction in data input and agent unreliability.
• I built a custom Slack inbox. It was easier than you think. | Yash Tekriwal (Clay) — AI can transform overwhelming digital communication into manageable, prioritized workflows through custom tools.
• Get back hours every day with autonomous agents in Amazon Quick — Autonomous AI agents and integrated activity feeds automate routine tasks and consolidate information, significantly reclaiming professional time.
• An AI app prepares me for my day now - and I've never been more organized — Huxe creates personalized daily audio briefings from user data and news, enhancing organization without screen time.
• I built an app for a teacher to save her time grading. — AI tools can effectively augment human expertise by proposing drafts for validation, significantly reducing repetitive tasks.
• I put Google’s 24/7 AI assistant Gemini Spark to work, and it’s actually pretty useful — Google's Gemini Spark demonstrates agentic AI's utility for automating personal and work-adjacent tasks, despite current integration and branding limitations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Towards Data Science.