Instant Answers from 100GB+ of Google Drive Documents
A RAG-powered chatbot that lets employees search Google Drive in plain language
About the customer
Our client is a US company whose employees often needed to find important files, such as PDFs, DOCX, PPTX and TXT files, among more than 100GB of documents on Google Drive. Searching by hand was slow and inefficient. We built a chatbot that lets employees ask questions about their Google Drive content in plain language, saving time and reducing manual searching.
What needed solving
100GB+ of documents
The client had over 100GB of corporate documents in many formats, including PDFs, spreadsheets and text files
Manual searching
Employees had to sift through the data manually, which reduced productivity
Natural-language access
Users needed to find information using natural language, without navigating Google Drive by hand
How we built it
Data loaders
Separate loaders for Google Docs, DOCX, PDFs, PPTX and TXT files read the data from Google Drive into a vector database, which stores text so it can be searched by meaning.
Retrieval
LlamaIndex splits documents into smaller chunks and finds the segments relevant to each question.
Answer generation
An OpenAI GPT model, integrated with LlamaIndex, writes accurate answers that fit the question.
Chat history
MongoDB stores user details and chat sessions, so users can return to their previous questions.
Deployment
Packaged with Docker as a Flask-based web app.
Employees find documents by asking questions, not by searching manually
Technologies
Domains
Have a Query? Let's Have a Conversation
Do you have a project in mind? We’d love to know what it is. Email us or send a message from our contact page.
Have a similar problem?
Tell us what you are trying to solve and our team will get back to you.