Portfolio Agent

Agent Description:

The Portfolio Agent is an intelligent investment assistant that evaluates a user's current portfolio holdings and offers data-driven recommendations for rebalancing. It combines performance analysis with real-time market intelligence to identify which assets to trim, hold, or increase for optimal portfolio health.

Purpose and Components
  • Purpose: To analyze user-submitted investment portfolios, assess asset performance, gather market insights, and generate actionable rebalancing recommendations.
  • Components:
    • A performance extractor to compute metrics and portfolio composition.
    • A rebalancing advisor that integrates performance with news data to guide investment decisions.
Supported Capabilities
  • Ingesting portfolio data from CSVs or documents.
  • Calculating asset-level and portfolio-level performance metrics.
  • Identifying over/underweight holdings.
  • Fetching real-time stock news, analyst sentiment, and sector trends.
  • Providing clear recommendations: Trim Position, Hold, or Consider Adding.
  • Generating user-friendly summary reports for investors.
LLM Used
  • Google Vertex
    Note: To learn more about the LLM and to modify its behavior, refer to the Configuring LLM settings section.

Sub-Agents

1. Performance Extractor Agent

  • Role:Data Extractor.
  • Scope: Process portfolio files to compute market values, gains/losses, and weight distribution.
  • Description: This subagent receives user-submitted portfolio data (for example, CSV or DOCX with columns like Ticker, Quantity, and Average Cost) and enriches it using real-time market prices fetched from financial APIs. It calculates:

    • Current Market Value

    • Unrealized Gain/Loss

    • Portfolio Weight (%)

      These metrics are passed to the Rebalancing Advisor for further analysis.

  • Tools used: GET Request Tool (for retrieving financial data and current market prices)
  • LLM Used: Google Vertex (inherits from parent).

2. Rebalancing Advisor

  • Role:Rebalancing Advisor.
  • Scope:Analyze portfolio performance, search for relevant news, and provide investment recommendations.
  • Description: This agent performs both performance analysis and market intelligence synthesis. It receives structured holdings and, for each significant asset (typically over 5% of the total portfolio), searches for the latest news, analyst ratings, and industry insights (for example, HDFC Bank stock outlook July 2025, Indian pharma sector 2025 trends). It uses this information to make specific recommendations:

    • Trim Position: For overweight holdings with negative outlook or risks.

    • Hold: For stable performers with neutral news sentiment.

    • Consider Adding: For underweight holdings with strong prospects.

  • A final summary report is generated with clear, actionable suggestions and reasoning for each recommendation.
  • LLM Used: Google Vertex (inherits from parent).
  • Tools used: Google Search Tool
Tools Used:
  • GET Request Tool (for market price and metrics)
  • Google Search Tool (for retrieving recent news and sentiment)
Note: For details on modifying the Tools, refer Tools Library section.
Agent Workflow Behavior Summary
  • The Performance Extractor Agent parses and computes performance metrics from the uploaded portfolio data.
  • The Rebalancing Advisor augments this analysis with external news and sentiment from the web, synthesizes findings, and provides clear rebalancing guidance.
  • The output is a comprehensive yet user-friendly report detailing gains, weightings, news impacts, and recommended asset adjustments.
Sample Questions:
  • Can you analyze my investment portfolio and suggest changes?
  • Which of my holdings are underperforming or overweight?