If the year 2023 is all about the rise of large language models (LLM), the year 2024 will be about retrieval-augmented generation (RAG) — a technology that grounds LLMs with your own data and knowledge. RAG not only enhances the personalization and reduces hallucination of AI-driven solutions, but also unlocks new levels of efficiency and innovation for businesses, propelling them ahead in the competitive landscape.
Today, we are super excited to announce the availability of our new Epsilla Cloud product– a Retrieval-Augmented Generation as a Service (RAGaaS) platform designed to democratize generative AI application building for all. With Epsilla Cloud, you can easily and quickly create personalizable AI Chatbots grounded with your own data from many sources (local files, cloud storage, Sharepoint, Google Drive, Dropbox, Notion, …), and pair it with LLM of your choice (OpenAI, Anthropic, Mistral AI, …).
Let me show you how simple it is to create your own AI chatbot below.
Example: Build my AI Tax Assistant in 5 minutes
Tax season is approaching, a period typically filled with stress and confusion. I’m wondering: can I develop a tax assistant to ease my anxiety? So I created a Tax Assistant using Epsilla Cloud, to answer my tax inquiries with answers grounded by uploaded IRS publications with reference to pertinent parts of the source documents.
Creating my tax assistant takes only 3 steps:
Step 1. Create my knowledge base
Switch to Knowledge Bases tab:
Then click Create Knowledge Base
Here we choose Local Files as the knowledge base type
For building the tax assistant, I am uploading a few of IRS publications:
You can check the processed data at the Data Storage tab:
Note: Epsilla Cloud provides customizable options for advanced data loading, chunking, and embedding to improve the quality of the semantic search index on your data. We will have a series of follow up articles to talk about it.
Step 2. Create the tax assistant chatbot
Now click the Create Application button at the navigation bar:
And choose Chatbot as the application type
I can add a name to my tax assistant, provide a description, and upload a logo:
My chatbot is up and running in a few seconds:
Now let me customize the chatbot role. Technically this is the system prompt to instruct the LLM to behave with desired personality:
You are a professional CPA specializing in individual taxation. Utilize your expert knowledge of tax laws and regulations to ensure clients benefit from all applicable deductions and credits. Provide clear explanations to clients about their tax situations and the implications of their financial decisions.
And add the knowledge from IRS documents we just uploaded.
Now let’s configure the look and feel of the chatbot, including the introduction message (first message the chatbot says for a new conversation), sample questions (for the chatbot end user to quickly start using), and enable follow up questions.
Now my Tax Assistant chatbot is ready to use. Give it a try:
As you can see, the chatbot not only gives very informative answers to my tax questions, but also provides reference links in the answer so I can check out relevant parts of the documents via highlighted text.
Note: Epsilla Cloud provides advanced configurations for improving the quality of the question answer of the chatbot, including prompt engineering, query rewriting, and advanced knowledge retrieval and reranking techniques. We will have a series of follow up articles to talk about them.
Step 3. Publish and Share
Once I am happy with the Tax Assistant, I can publish the chatbot and share with my friends, family and colleagues by allowing all Epsilla Cloud users to access:
Then any Epsilla Cloud user with the access link can chat with my chatbot:
You can try my tax assistant to see if that is helpful to address your tax questions. Please let us know how it works. (Disclaimer: the knowledge used to augment LLM in this demo is from a few IRS documents for demo purpose; please consult a tax professional for the accuracy of the answer)
Conclusion
Epsilla Cloud drastically simplifies the complex process of data integration and chatbot development, making it accessible to anyone, regardless of their technical expertise. Whether you’re looking to enhance customer service, streamline internal communications, or explore new interactive experiences, Epsilla Cloud offers a user-friendly, flexible, and efficient solution. By leveraging the power of Epsilla Cloud, users can unlock the full potential of their data, transforming it into intelligent, conversational agents that drive engagement and deliver insights. Join us in exploring the future of AI interactions with Epsilla Cloud.