Introduction
InstaStack is an innovative Generative AI feature of Botco.ai that allows users to enrich their chatbot's knowledge base with a multitude of resources. This feature is pivotal for integrating customer data, including documents and files that are not publicly available on the website, ensuring that the chatbot can provide comprehensive responses to user inquiries. This is our take on Retrieval Augmented Generation (RAG).
Overview
InstaStack is divided into two main components: the Knowledge Base and the AI Playground. The Knowledge Base is where users upload valuable data that can range from internal documents to sensitive PDFs and HTML files. The AI Playground is a testing ground where users can query the uploaded data to evaluate the chatbot’s response accuracy.
Features
Knowledge Base (Docs + Crawler)
- Docs: Users can upload documents to the InstaStack Knowledge Base. These documents can be PDFs, HTML files, or any data typically stored in a cloud service like Google Drive.
- Crawler: The system includes a dedicated crawler that indexes a customer’s website content separately, so manual addition of web pages is unnecessary. Users can choose to extract all pages under a given URL or directly from a sitemap. The content from these URLs is then indexed and added to the knowledge base. You can also schedule a crawl at the frequency of your choosing (daily, weekly, monthly, quarterly, yearly) to ensure the data is up to date.
- Collections: Organize documents into collections based on different topics or departments. This allows for targeted responses from the chatbot based on the user's current page or inquiry topic.
AI Playground
- Testing Queries: After uploading documents to a collection, users can test the effectiveness of the retriever by asking relevant questions in the AI Playground.
- Prompt Customization: The prompt paired with the user input can be modified to fetch the most appropriate answer from the collection.
- Retriever Accuracy: The AI Playground shows how well the retriever can pull relevant data from the collection, displaying the document and passage where the answer was sourced.
- Iteration: Customers use this tool to iterate on their chatbot’s responses, ensuring the data and formatting are optimized to answer potential questions accurately.
Conclusion
InstaStack empowers users to significantly enhance their chatbot’s understanding and response capability by feeding it a rich diet of relevant information. With the ability to organize and test data, users can ensure their chatbot provides precise and helpful information, thereby improving the overall user experience and efficiency of the chatbot.
FAQ Section
Q: What types of documents can be uploaded to InstaStack? A: You can upload various types of documents, including PDFs, HTML files, and text documents.
Q: How does the InstaStack crawler differ from manual uploads? A: The crawler automatically indexes website content, while manual uploads are for documents not available on the website.
Q: Can I restrict the chatbot to use data from a specific collection? A: Yes, you can assign chatbot queries to specific collections to ensure it uses relevant data.
Q: Is there a limit to the number of documents I can upload to InstaStack? A: Document upload limits may depend on your subscription plan with Bot CoAi.
Q: How secure is data uploaded to InstaStack? A: InstaStack employs robust security measures to ensure your data is protected and complies with privacy regulations.
Q: Can I delete documents from InstaStack if they are outdated? A: Yes, users can manage their documents and delete outdated files as needed.
Q: What if the AI Playground suggests an inaccurate response? A: Users can provide feedback and train the AI model by editing the response, which helps improve future interactions.
For more detailed instructions on how to leverage InstaStack for your chatbot, or if you encounter any issues while using it, our customer support team is ready to assist you every step of the way.