Agent for Personalized Market Insights with Voice UI
Develop a sophisticated financial analyst agent using the OpenAI Agents SDK that provides personalized market insights. This agent will track specific company or sector news. It will utilize GPT-5 for advanced reasoning, integrate with Alpha Vantage for real-time financial data, and provide interactive insights via a voice interface powered by Resemble AI. The agent should be capable of managing personalized investment portfolios and providing proactive updates. TPOT can be integrated for automated data processing or model evaluation workflows.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an OpenAI agent for personalized market insights, integrating real-time financial data and a voice interface.
How work is evaluated
The agent will be evaluated on its ability to accurately retrieve and synthesize financial information, provide relevant personalized insights, and respond effectively through a voice interface.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Valid JSON Output
Checks if the JSON summary is correctly formatted and non-empty.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Relevant Information
Verifies that the response contains information related to the query's core topics.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Accuracy of Financial Data
Measures the correctness of retrieved financial data (e.g., stock prices, if available). • target: 90 • range: 0-100
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Response Personalization Score
Evaluates how well the agent tailors information to tracked companies or user history. • target: 85 • range: 0-100
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Master the OpenAI Agents SDK for defining tools, function calling, and managing agent conversations.
Integrate GPT-5 as the core reasoning engine, leveraging its multimodal capabilities for richer data interpretation (e.g., analyzing chart data).
Utilize the Alpha Vantage API to fetch real-time stock quotes, company financials, and market news programmatically.
Implement a voice-enabled interaction layer using Resemble AI for text-to-speech and speech-to-text functionalities.
Design a system for persistent memory to store user preferences and track specified investments over time.
Automate data preparation and validation pipelines using TPOT for optimizing financial data processing before agent ingestion.
Build custom tools for the OpenAI agent to perform actions like 'get_stock_price', 'get_company_news', and 'track_portfolio'.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
The agent will be evaluated on its ability to accurately retrieve and synthesize financial information, provide relevant personalized insights, and respond effectively through a voice interface.
Challenge input
{ "user_query": "string", "tracked_companies": ["string"] }
Agent execution
No agent tool is listed for this challenge.
Evaluated output
{ "voice_response_transcript": "string", "json_summary": { "company": "string", "latest_price": "float", "news_summary": "string", "analyst_sentime...
- Checks if the JSON summary is correctly formatted and non-empty.
- Verifies that the response contains information related to the query's core topics.
- Measures the correctness of retrieved financial data (e.g., stock prices, if available). • target: 90 • ran...
- Accuracy Of Financial Data target: 90
- 1 public reference case
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Valid JSON Output · Weight 1
- Relevant Information · Weight 1
- Accuracy of Financial Data · Weight 1
- Response Personalization Score · Weight 1
Metrics
- Accuracy of Financial Data · Target 90
- Response Personalization Score · Target 85
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: { "user_query": "string", "tracked_companies": ["string"] }
Output: { "voice_response_transcript": "string", "json_summary": { "company": "string", "latest_price": "float", "news_summary": "string", "analyst_sentiment": "string" } }
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.