Back to Projects

TV Stock Terminal

A family-oriented securities terminal for TV screens, integrating voice recognition and AI assistant capabilities to provide stock market information, simulated trading, and investment education.

Finance Voice AI Smart TV LLM

Project Background

Status

In Progress

Year

2024

Duration

7 months

Category

Television Application

Company

China Telecom

My Role

Wannz Product Director
TV Stock Terminal

A family-oriented securities terminal for TV screens, integrating voice recognition and AI assistant capabilities to provide stock market information, simulated trading, and investment education.

TV Stock Terminal: Building a Comprehensive Home Investment Experience

Project Overview

The TV Stock Terminal project represents a strategic collaboration with a major telecommunications operator to transform traditional set-top box devices into sophisticated financial terminals for home users. By leveraging the large screen experience of television and incorporating advanced AI capabilities, we created an accessible platform that democratizes stock market information and investment education.

Strategic Challenges & Solutions

Bridging the Digital Divide in Investment Tools

Challenge: Investment tools have traditionally been confined to mobile apps and professional desktop terminals, creating barriers for older users and those seeking a more comfortable viewing experience.

Solution: We repurposed the ubiquitous set-top box infrastructure to deliver financial information on the largest screen in most homes, creating an inclusive experience that entire families could engage with.

Technical Integration Complexity

Challenge: Integrating voice recognition, real-time market data, and AI assistants into set-top box hardware with limited processing capabilities.

Solution: Developed a lightweight client architecture with cloud-based processing for voice recognition and AI functions, enabling smooth performance even on older set-top box models.

Key Features & Innovations

Intuitive Voice-Controlled Interface

  • Implemented natural language processing allowing users to request stock information through simple voice commands like "Show me today's performance of Tencent stock"
  • Created contextual awareness so users can follow up with questions like "How about its competitors?" without specifying the industry again

Comprehensive Financial Information Center

  • Real-time stock market data with customizable watchlists and alerts
  • Technical analysis tools optimized for TV display, with clear visualization of trends
  • Industry news and updates with AI-summarized highlights to quickly grasp key information

Interactive Learning Environment

  • Educational video library covering investment fundamentals to advanced strategies
  • Interactive simulated trading platform where users can practice investment strategies without financial risk
  • Step-by-step guided tutorials for new investors with voice-interactive Q&A

AI Investment Assistant

  • Integrated large language model capabilities to answer investment-related questions
  • Personalized investment education pathways based on user knowledge level
  • Explanation of complex financial terms and concepts in accessible language

Technical Implementation

Architecture Overview

The solution follows a hybrid architecture model:

  • Lightweight TV client application optimized for set-top box hardware
  • Cloud-based voice processing and AI computing
  • Real-time data streaming for market information
  • Secure API integration with financial data providers

Voice Recognition Optimization

One of the core technical challenges was optimizing voice recognition for the living room environment:

  • Created a financial-specific language model to improve recognition accuracy for stock names and investment terminology
  • Designed fallback mechanisms when voice commands are ambiguous, presenting visual options for clarification

Large Model Integration

We developed a specialized financial knowledge version of the large language model:

  • Fine-tuned on financial education materials and market analysis
  • Optimized response generation for TV display (concise yet informative)
  • Implemented strict guardrails to prevent providing specific investment advice or recommendations

User Experience Design

Accessibility-First Approach

  • Large, readable fonts and high-contrast visuals suitable for viewing from a distance
  • Voice-control alternatives for all navigation functions
  • Multi-generational testing to ensure usability across different age groups and technical proficiency levels

Engagement Mechanics

  • Daily market summaries and personalized watchlist updates
  • Gamified investment simulation with progress tracking
  • Seasonal investment education campaigns aligned with market events

Business Impact & Future Development

Current Results

  • Successful deployment across 2,000+ set-top boxes in the initial phase
  • Average user engagement of 15 minutes per session

Roadmap

  • Family account features allowing personalized experiences for different household members
  • Integration with actual brokerage services for seamless transition from education to real investment
  • Expanded AI assistant capabilities for more sophisticated market analysis and educational content

Reflections & Lessons Learned

This project highlighted the untapped potential of television as a platform for financial services. While mobile will remain dominant for transactions, the comfortable viewing experience and family-oriented nature of TV creates unique opportunities for financial education and information consumption.

The careful balance between technical sophistication and accessibility proved crucial to the project's success. By focusing on solving specific pain points for home users rather than replicating professional trading terminals, we created a genuinely valuable new touchpoint in the financial ecosystem.

(Note: The data involved in this article has been desensitized or obfuscated, and some technical details have not been disclosed due to confidentiality agreements)