Introducing Koo Bot: Your AI-Driven Document Parser

Want to save time at work, instead of spending hours reading documents to find a 2 line answer to something? Worry not! Here’s Koo Bot for you - developed by TheCodeWork®.

Koo Bot is an AI driven tool that revolutionalises how companies work with documents through Slack(for now - Can be customised for other platforms as well). It is capable of reading through an entire document that is uploaded or shared on slack and then revert with quick and precise answers to all the questions that may be asked about the document.

It’s time to stop scrolling and searching through lengthy reports – meet Koo Bot for productivity and saving your time.

How Koo Bot Works ?

Unlock the Power of Automation: Effortlessly Analyze, Extract, and Manage Documents with Koobot's Advanced AI Parsing Technology.

Seamless Document Parsing

Once you or anyone else, uploads a document on Slack, Koo Bot immediately starts to process the information, after you have added the tool on that specific Slack channel.

Intelligent Question Answering

Type any question concerning the document and Koo Bot provides the answer by summarising the respective section to enable you make the search within a few seconds.

Quick and Accurate

With help of Koo Bot you get exact and relevant replies whether it is a full report or a brief summing up thus saving a lot of time on documents reading.

Why Choose Koo Bot?

Save Time

Once you or anyone else, uploads a document on Slack, Koo Bot immediately starts to process the information, after you have added the tool on that specific Slack channel.

Boost Productivity

Relieve yourself from the stress of searching for documents. Work on the other priorities instead!

Enhance Collaboration

Everyone has easy access to the required information and the processes of decision-making will be faster.

Technologies Used

Slack Events API

Listens for events like file uploads or user messages in a Slack workspace, triggering the chatbot to process documents and respond.
Managed using the Python slack_bolt and slack_sdk libraries.

Flask

Acts as the web framework to handle Slack's webhook requests and route them to the appropriate functions.
Manages the backend logic, such as document processing, embedding generation, and query handling with Pinecone and GPT-3.5.

Document Processing

Uses libraries like python-docx for .docx files, pdfplumber for .pdf, and striprtf for .rtf to extract text.
Text is parsed from uploaded documents and passed to the language model for further processing.

Pinecone

Stores vector embeddings of documents generated by GPT-3.5 for efficient retrieval during queries.
When a user queries the chatbot, it searches Pinecone for relevant documents using vector-based similarity search.

GPT-3.5 (With RAG)

Generates embeddings from document text and retrieves relevant content from Pinecone.
GPT-3.5 uses the retrieved documents to create intelligent, context-aware responses to user queries.

Langchain

Manages the entire process of chaining tasks, such as document retrieval and GPT-3.5 response generation.
Facilitates the Retrieval-Augmented Generation (RAG) workflow, efficiently handling the interaction between Pinecone and GPT-3.5.

Workflow summary

Install the app into Slack workspace and invite it to the desired channel where it will listen for events.

Slack event triggers Flask via a webhook.

Flask extracts text from uploaded documents.

GPT-3.5 generates embeddings and stores them in Pinecone.

For user queries, Flask retrieves relevant documents from Pinecone using Langchain.

GPT-3.5 generates a context-based response, which Flask sends back to Slack.

Upcoming Features - to look out for!

Integrate your own LLM.

Integrate your own pinecone account.

Customize your messages.

Who Can Benefit?

Koo Bot is perfect for teams across industries.