← Back to Portfolio
AutoAnsweringForm — AI Automated Q&A Agent
An AI-powered automated question-and-answer system designed to assist users in completing structured Q&A forms using knowledge retrieved from both internal databases and the internet.
Inspiration
Many modern workflows require answering structured forms quickly and accurately. These include job recruitment tests, research questionnaires, knowledge competitions, and corporate knowledge documentation.
In many situations, users must search the internet manually, verify information, and rewrite answers into form fields. This process is repetitive and time-consuming.
The idea behind AutoAnsweringForm was to create an AI agent capable of performing these steps automatically:
- Search for information from multiple sources
- Retrieve relevant knowledge from internal databases
- Combine internal and external data
- Generate clear answers automatically
- Fill Q&A responses faster and more efficiently
The system acts as a knowledge retrieval assistant capable of supporting continuous research and automated answer generation.
What It Does
- Automatically searches the internet for relevant information
- Integrates corporate or internal knowledge bases
- Generates AI-based answers for structured Q&A forms
- Reduces manual research time
- Supports continuous knowledge retrieval
- Creates consistent answers across multiple users
The platform can operate in multiple search modes:
- Internet-only mode — answers retrieved from external sources
- Internal knowledge mode — answers generated from corporate databases
- Hybrid mode — combines internal knowledge with external web data
Use Case Scenarios
HR Recruitment Testing
Human Resources departments sometimes conduct Q&A based recruitment tests.
This platform can help generate structured answers automatically while ensuring each candidate receives unique responses generated from fresh knowledge sources.
Urgent Q&A Completion
Some users must complete large sets of questions quickly. AutoAnsweringForm accelerates this process by generating answers instantly using AI retrieval systems.
Corporate Knowledge Base Enrichment
Organizations can populate internal knowledge bases using scraped data and structured documents.
Once stored, employees can retrieve answers instantly without repeating research tasks.
Research and Learning Support
Students or researchers can use the system as a learning tool that continuously retrieves new knowledge from the internet while referencing stored knowledge archives.
How It Was Built
The system follows a distributed microservice architecture designed to handle knowledge ingestion, processing, embedding generation, and AI retrieval.
- Browser automation services scrape and collect information
- Text chunking services prepare knowledge for AI processing
- Vector embedding generation transforms text into machine-searchable vectors
- Vector search enables semantic similarity retrieval
- Streaming infrastructure distributes data between services
- Multiple workers run in parallel to scale scraping and data processing
This architecture allows the platform to continuously expand its knowledge base while maintaining fast query performance.
System Architecture
- Frontend Q&A form interface
- Scraping workers collecting internet knowledge
- Chunk generator preparing documents for AI models
- Vector embedding generator creating searchable embeddings
- Vector database storing semantic knowledge
- AI retrieval engine generating answers
- Streaming data pipeline connecting services
Each component operates independently, allowing horizontal scaling when data collection or answer generation demand increases.
Example Workflow
User submits a question.
System checks:
1. Internal knowledge base
2. Scraped external knowledge
3. Vector similarity search
Relevant information is retrieved.
AI generates a clean and structured answer.
Question:
"What are the latest cybersecurity trends?"
System retrieves:
• Internal research reports
• Latest web articles
• Security blog content
AI summarizes and returns a concise answer.
Challenges
- Handling large volumes of scraped knowledge
- Maintaining answer accuracy across multiple sources
- Managing distributed worker coordination
- Ensuring vector search relevance
- Balancing performance with infrastructure cost
Accomplishments
Built a distributed AI knowledge pipeline capable of collecting external information, processing documents, generating embeddings, and answering questions automatically through a scalable architecture.
The system demonstrates how AI-assisted knowledge retrieval can dramatically reduce manual research and improve productivity when handling structured question workflows.
Technology Used
- TiDB
- Python
- REST API
- ReactJS
- Node.js
- Kafka
- Distributed Puppeteer
- HTML
- CSS
- JavaScript
- AI RAG Langchain
- Vector Indexing
- Custom Smart ReactJS Components
- Microservices