Thoughtful AI Customer Support Agent

A simple conversational agent that answers questions about Thoughtful AI's agents (EVA, CAM, PHIL) using a predefined dataset, with a friendly fallback for unrelated questions.

Features

• Conversational UI: Web-based chat interface (Gradio).
• Predefined answers: Matches user questions to a hardcoded QA set in temp.json.
• Similarity matching: Uses fuzzy matching so rephrased questions still get the right answer.
• Fallback: When no good match is found, responds with a helpful message and suggests related topics.
• Error handling: Handles empty input, missing dataset, and invalid JSON gracefully.

Setup

1. Python: Use Python 3.10+ (for list[dict] type hints; works on 3.9 with from __future__ import annotations if needed).

2. Install dependencies:

   pip install -r requirements.txt

3. Dataset: Ensure temp.json is in the same directory as agent.py (included in this repo).

Run the app

python app.py

Then open the URL shown in the terminal (e.g. http://127.0.0.1:7860) in your browser.

Project structure

• agent.py – Core logic: load QA dataset, similarity matching, get_answer() with fallback.
• app.py – Gradio chat UI and entry point.
• temp.json – Predefined questions and answers about Thoughtful AI.
• requirements.txt – Python dependencies.

Usage tips

• Try questions like: "What does EVA do?", "Tell me about Thoughtful AI's Agents", "How does PHIL work?"
• Rephrased or partial questions (e.g. "EVA?" or "benefits of your agents") are matched when similar enough to the dataset.
• Off-topic questions receive the fallback message guiding users back to agent-related topics.

Evaluation notes

• Functionality: Conversation flow and retrieval from the dataset are implemented; fallback is used when no match is found.
• Code quality: Single-purpose functions, type hints, and a clear split between agent logic and UI.
• Robustness: Empty input, missing/corrupt temp.json, and unexpected errors in get_answer are handled with clear user-facing messages.