Project Stats

Name
Investing.com
Industry
Finance
Employees
250+
Location
Tel Aviv
Year
2023
Duration
1 year and ongoing
Team
1 AI Engineer, 1 Full Stack Engineer
TechStack
  • OpenAI GPT 3.5, 3.5 Turbo, 4
  • Google PaLM, Gemini
  • PostgreSQL, Pinecone
  • React, NextJS
  • Python, LangChain, FastAPI, Gradio
KeyResults
  • Built MVP in 2 months
  • Built in-house playground app for internal testing and prompt engineering
  • Scaled system to handle 10K concurrent users
  • Successfully soft-launched to InvestingPro users in 9 months
Investing Company Logo

Investing.com | Developing an AI Chatbot for Investors

Investing.com

Investing.com is a financial markets platform serving over 46 million monthly users in 44 languages. It offers real-time quotes, charts, financial news, technical analysis, and tools covering stocks, commodities, cryptocurrencies, bonds, and forex, supporting informed investment decisions worldwide.

[@portabletext/react] Unknown block type "image", specify a component for it in the `components.types` prop

Project Overview

The client recognized the need for an intelligent solution to help their premium subscribers navigate the complex landscape of financial data. They wanted a versatile, user-friendly assistant that could understand and respond to a wide range of financial queries with accuracy and insight.

Goals

To create an AI chatbot capable of:

  1. Interfacing with their internal financial query API to access metrics for over 70,000 financial entities
  2. Screening stocks based on various financial parameters
  3. Retrieving and synthesizing relevant financial news, analyst reports, and earnings call transcriptions
  4. Handling complex queries that require data from multiple sources
  5. Scaling to support thousands of concurrent users

Challenges

Text-to-FinAPI Translation

The FinAPI provides access to thousands of metrics for more than 70,000 financial entities. Translating natural language queries into precise API calls required a sophisticated approach beyond simple prompt engineering. Simply passing the API documentation as an LLM prompt is not possible.

Financial Expertise in Prompt Engineering

While AI developers could handle basic prompt engineering, crafting high-quality responses required input from financial experts who could interpret complex financial data and provide meaningful insights.

Multi-Source Query Handling

Many user queries required information from multiple sources, necessitating a system that could dynamically aggregate data from financial metrics, news articles, and other relevant sources.

Scalability and Performance

The system needed to handle a high volume of concurrent users without compromising on speed or accuracy.

Solutions

LLM-Powered Entity Extraction System

We developed an entity extraction system to identify relevant financial entities and metrics from user queries, enabling accurate data retrieval from the FinAPI.

Playground Gradio App

To streamline development and testing, we built a custom Gradio app that enabled non-technical team members to easily test and refine chatbot responses. This tool provided comprehensive control over LLM selection, parameters, and system prompts, supporting rapid iterations and improvements.

Dynamic RAG System

For handling complex, multi-source queries, we developed a dynamic Retrieval-Augmented Generation (RAG) system. This approach analyzed user queries to determine the appropriate data sources, retrieved the necessary information, and combined it for the LLM to generate detailed, accurate responses.

Scalable Cloud Infrastructure

We designed and deployed a robust, cloud-based infrastructure to handle high demand. This included optimized database queries, caching strategies, load balancing, and auto-scaling capabilities, ensuring consistent and reliable performance during peak usage.

Results

Increased Conversion Opportunities

In a targeted user feedback survey, 70% of non-premium users showed interest or willingness to upgrade to a premium account to continue using the AI chatbot, with 42% expressing a strong likelihood of doing so.

Rapid MVP Development

We successfully delivered a functional Minimum Viable Product (MVP) in just two months, showcasing our ability to quickly transform concepts into working solutions.

Efficient Internal Testing and Iteration

Our custom-built Gradio-based playground app empowered non-technical team members to test and refine the chatbot's responses, significantly enhancing the quality and accuracy of AI-generated financial insights.

Scalable Infrastructure

Successfully scaled the system to handle thousands of concurrent users, ensuring robust performance and positioning the product for future growth.

Timely Product Launch

Soft-launched the product to InvestingPro users within nine months, allowing for real-world testing and user feedback collection.

Enhanced User Experience

The AI assistant significantly improved user engagement and satisfaction by providing quick, accurate responses to complex financial queries.

By leveraging cutting-edge AI technologies and financial expertise, this project delivered a powerful, scalable solution that revolutionized how Investing.com's users interact with financial data.

Ready to grow your idea 10x?

MVP, POC, production-grade pipelines, we have built it all

No salespeople, no commitment