-
-
Explore financial decisions before you commit.
Inspiration
What if you could test a financial decision before making it?
Taking a car loan or making a major purchase affects more than your monthly budget. It can change your savings, financial flexibility, and ability to meet future goals. I built Agentic Financial Planning & Simulation to help people explore these trade-offs before committing.
What It Does
The project is a conversational AI financial-planning agent that turns natural-language requests into financial actions, calculations, and simulations.
Its capabilities include:
- Financial Profile Management: Retrieve and update the financial profile, including income, expenses, savings, and goals.
- Expense Management: Add new expenses to evaluate their impact on the financial plan.
- Financial Event Management: Add, update, and remove planned financial events, such as recurring EMIs or one-time purchases.
- Scenario Simulation: Simulate the current financial plan or run a separate hypothetical scenario over multiple months.
- Scenario Comparison: Compare alternative financial decisions and understand their different effects on cash flow and savings.
- Find My Limit: Calculate the maximum additional monthly commitment that satisfies the configured savings goal and minimum savings-buffer requirements.
The 10 MCP Tools
The agent uses 10 tools through the Model Context Protocol (MCP):
get_profile— Retrieve the current financial profile.set_profile— Update the financial profile.add_expense— Add an expense to the plan.add_event— Add a planned financial event.update_event— Modify an existing financial event.remove_event— Remove a financial event.simulate_current_scenario— Simulate the current saved plan.simulate— Run a financial simulation for a specified hypothetical scenario.compare— Compare multiple financial scenarios.find_limit— Calculate an affordable additional monthly commitment under the configured financial constraints.
Together, these tools allow the agent to do more than answer questions: it can retrieve information, manage planned changes, run calculations, and evaluate alternative futures.
How I Built It
I built the application using Python and Streamlit. An OpenAI model interprets natural-language requests and coordinates the MCP tools, while Python-based financial logic performs the underlying calculations and simulations.
A key design goal was to keep hypothetical scenarios separate from the saved financial plan. This allows users to explore decisions without unintentionally changing their existing plan.
The project is presented as a simulated Alexa+ conversational experience through a custom web interface. It does not claim direct integration with an Alexa device or the Alexa+ service.
Challenges and Learning
One challenge was ensuring that scenario comparisons used the correct assumptions and produced consistent results. I tested different EMI amounts, checked savings-goal and minimum-buffer constraints, and verified that hypothetical simulations did not unintentionally modify the saved plan.
This project helped me understand how to connect an LLM to multiple tools, coordinate an agentic workflow, and combine natural-language interaction with deterministic financial calculations. It also reinforced the importance of validating tool outputs rather than relying on the language model alone for arithmetic.
Why It Matters
Instead of simply asking whether a purchase is affordable today, users can explore how it could affect their finances over time. The goal is to make financial trade-offs easier to understand before making a commitment.
Built With
Python, Streamlit, OpenAI API, Model Context Protocol (MCP).
Disclaimer
This is an experimental financial-planning assistant that uses a sample financial profile. Its outputs are estimates for educational purposes and are not professional financial advice.
Built With
- agentic
- ai
- api
- context
- financial
- generative
- model
- openai
- planning
- protocol
- python
- streamlit
