💼 AI Investor Assistant API – Powering the Next Generation of Financial Intelligence

🌟 Overview

The AI Investor Assistant API is a plug-and-play SaaS backend that empowers any financial product—be it an investment app, trading platform, portfolio tracker, or finance education tool—to offer intelligent, personalized investor guidance via natural language chat.

Built to seamlessly integrate into existing systems, this API delivers deep financial insights by combining:

• Real-time user portfolio data,
• Sector & stock-level context,
• State-of-the-art LLM technology.

Whether you’re looking to engage users, reduce churn, increase retention, or stand out with AI-driven features, this toolkit gives your app a serious competitive edge—with zero AI infrastructure overhead.

🚀 Why It Matters for Your App

🧠 Transform User Experience

Engage your users with smart, context-aware financial conversations—just like having a personal financial assistant inside your app.

📊 Personalization That Scales

Tailor every answer based on each user’s risk tolerance, age, portfolio composition, and investing goals. Go beyond generic responses.

🔁 Increase Engagement & Retention

Use automated follow-up questions, personalized investment education, and FAQ discovery to keep users exploring and coming back.

💸 Save Time & Resources

Skip the costly R&D of building your own AI stack. With this API, you get a prebuilt, production-ready AI assistant tailored for finance.

💡 What Can It Do?

Here’s what your app unlocks with the AI Investor Assistant:

🧱 How It Works – Under the Hood

The system is composed of modular, interoperable endpoints that handle everything from session management to chat to follow-up generation.

1. Start a Session

Every conversation starts with:

POST /session

Returns a session_id used across endpoints to track chat history.

2. Optional: Provide User Context

Add user-specific data like portfolio and profile (age, risk level, etc.):

POST /user_context

Now all replies will adapt to this context. The assistant knows what the user holds and who they are.

3. Extract the Topic & Financial Tags

Not sure how to interpret a user query like:

“How did Microsoft and Apple perform in the tech sector?”

Let the API figure it out:

POST /chat/extract_topic_and_tags

Returns:

{
  "topic": "stock_overview",
  "topic_tags": {
    "stock_symbols": ["AAPL", "MSFT"],
    "sector_name": "Technology"
  }
}

4. Generate the Response

Send the enriched request to the main chat endpoint:

POST /chat

The assistant will stream back a detailed, human-like response incorporating everything it knows—context, holdings, market data, and tone.

5. Drive Engagement With Follow-Ups

Want to keep the user engaged and exploring?

POST /follow_up_questions

Returns a list of smart, personalized follow-up questions like:

• “Would you like to compare Microsoft’s and Apple’s R\&D spend?”
• “Should I include ETF alternatives in this analysis?”

6. Provide Learning Resources

Return beginner-friendly educational content by topic:

GET /faq?faq_topic=income_statement

Great for onboarding new investors and reinforcing trust in your platform.

🧪 Developer-First Example Workflow

# Create a session
POST /session

# Send a user query
POST /chat/extract_topic_and_tags
{
  "question": "Tell me about Apple and Microsoft’s financials.",
  "session_id": "abc123xyz"
}

# Use topic + tags to fetch the response
POST /chat
{
  "question": "Tell me about Apple and Microsoft’s financials.",
  "topic": "stock_overview",
  "session_id": "abc123xyz",
  "topic_tags": {
    "stock_symbols": ["AAPL", "MSFT"]
  }
}

🛠 Configuration and Setup

Supports both OpenAI and Ollama as model providers. Just set the .env like so:

LLM_PROVIDER=OPEN_AI
OPEN_AI_API_KEY=your-key
OPEN_AI_MODEL_NAME=gpt-4o-mini

# or for local LLMs via Ollama
LLM_PROVIDER=OLLAMA
OLLAMA_BASE_URL=http://localhost:11434
OLLAMA_MODEL_NAME=llama3.2

Other tunable settings include:

• FAQ_LIMIT: Number of FAQ items per topic
• CONV_MSG_LIMIT: How many past messages to retain in chat memory
• BASE_LLM_TEMPERATURE: Controls creativity vs. precision

🔍 How Topic/Tag Extraction Works

The system performs this in two stages:

1. Topic Extraction: LLM analyzes the user’s question to identify the domain (e.g. “stock_overview”).
2. Tag Extraction: Based on the topic, a separate prompt extracts relevant symbols, sectors, or financial docs.

✅ Modular and accurate ⚠️ Slightly more expensive due to two LLM calls (unless topic requires no tags)

See topic_tag_extractor.md for deep details.

🌍 Real-World Use Cases

💸 Investment Platforms

Add a built-in assistant that answers portfolio questions, explains stock movement, and suggests strategic rebalancing.

📊 Portfolio Apps

Offer daily summaries, risk analysis, or chat-based insights on held positions.

🏫 Financial Literacy Tools

Deliver educational Q\&A, contextual FAQs, and guided learning for beginners.

📈 Market News Aggregators

Transform headlines into plain-English explanations with sentiment and stock impact breakdowns.

🧩 Available API Endpoints

📈 Build Smarter Finance Experiences – Today

This API isn’t just a chatbot—it’s a financial brain for your application.

Instead of answering generically, it knows:

• Who the user is
• What they hold
• What they want to know
• What’s going on in the market

And it responds accordingly.

👩‍💻 Who Is It For?

• 🏦 Fintech Startups
• 📱 App Developers
• 💰 Wealth Management Platforms
• 📰 Market Analysis Tools
• 📚 Financial Education Portals

If your product has users that care about money, this assistant can help you build trust, engagement, and retention—while future-proofing your UX with AI.

📞 Contact

Reach out via email: stefanouorestis@gmail.com

Orestis Stefanou

Machine Learning Engineer, currently working at Plum Fintech