Workday Planner AI Agent

What It Does

Modern flat organizations face a critical challenge: as teams grow, maintaining context across all meetings and projects becomes overwhelming. Team members waste precious hours researching attendees, gathering background information, and understanding project contexts before important discussions.

Vertical-Org solves this by providing:

• Intelligent Meeting Preparation: Automatically researches attendees, topics, and relevant context
• Comprehensive Daily Briefings: Synthesizes calendar data with internal and external intelligence
• Context-Aware Research: Determines what type of research is needed based on meeting characteristics
• Actionable Insights: Delivers structured, immediately useful information for better decision-making

How We Built It

We built a sophisticated multi-agent orchestration system using Google's Agent Development Kit (ADK) that coordinates specialized AI agents to analyze different aspects of workday preparation:

Core Agent Architecture

Workday Planner Coordinator - Main orchestration agent that processes calendar data and delegates research tasks Calendar Parser Agent - Extracts and structures meeting information from calendar input Research Coordinator - Intelligently determines research strategy based on meeting context Web Search Agent - Conducts external research using BrightData MCP for public information Internal Research Agent - Searches internal systems (Gmail, Slack, JIRA) using Arcade tools Report Synthesizer - Compiles all research into actionable daily briefings

Smart Research Delegation

Our system intelligently routes research based on meeting characteristics:

• External meetings → Web research for attendee backgrounds and industry context
• Internal meetings → Internal system searches for project status and decisions
• Hybrid meetings → Combined research for comprehensive context

Technology Integration

Multi-Agent Coordination: Google ADK framework orchestrates specialized agents working in parallel External Intelligence: BrightData MCP integration for comprehensive web research and LinkedIn searches Internal Intelligence: Arcade tools integration for Gmail, Slack, and JIRA data access Model Flexibility: LiteLLM integration supporting Claude 3.5 Sonnet for high-quality analysis

Challenges We Overcame

Agent Coordination: Designed intelligent delegation system that prevents research duplication while ensuring comprehensive coverage Context Management: Built session-based memory isolation to maintain separate contexts for different workdays and meetings Real-time Integration: Seamlessly integrated multiple external APIs (BrightData, Arcade) for live data access Intelligent Routing: Created sophisticated logic to determine optimal research strategy based on meeting characteristics

Tech Stack

• Agent Framework: Google Agent Development Kit (ADK)
• Language Models: Claude 3.5 Sonnet via LiteLLM
• External Intelligence: BrightData MCP for web research and LinkedIn data
• Internal Intelligence: Arcade tools for Gmail, Slack, JIRA integration
• Configuration: Python-based modular configuration system
• Orchestration: Multi-agent coordination with intelligent delegation

Built With

• arcade.dev
• brightdata
• google-adk