CompaChat is a cutting-edge result-oriented communication platform designed to streamline your business interactions within your organization and with your customers. Whether you are looking to implement SOPs, unlock your trove of internal documents for employees, or convert customer queries into increased sales, CompaChat delivers focused responses that guide your audience into action.
Connect, collaborate, and celebrate the future of communication today!
Connect Anywhere, Anytime
Dive into a world where communication knows no bounds. CompaChat’s front-end connectors bridge the gap between you and your audience through WhatsApp, Facebook Messenger, SMS, and more. Stay connected, stay ahead!
Uncompromised Security
CompaChat’s industry strength security ensures that sensitive information remains confidential, accessible only to authorized personnel within your organization. With Role-Based Security, CompaChat safeguards your conversations, ensuring that confidentiality is never compromised. Secure, reliable, and ready for business.
Innovative at Its Core
CompaChat’s robust RAG architecture not only increases accuracy but also integrates seamlessly with the Role-Based Security, providing a secure and reliable platform.
Smart, Responsive, Result Oriented AI
Powered by Azure AI Search and Azure OpenAI LLM, CompaChat delivers intelligent, natural language responses tailored to user prompts and context. Focused responses delivered on actionable adaptive cards with live links, guide your audience and deliver results you’ve been looking for.
Responsible AI
Checkups on both incoming messages and generated responses ensure that they comply with your company policies and meet evolving regulations. CompaChat stands for ethical use of AI, ensuring a positive impact on society.
Robust Architecture
At the heart of CompaChat sit the RAG architecture, advanced AI Algorithms and Open AI LLM to bring you accuracy, security and actionable responses. Embrace the power of a platform built to deliver precision, protection and results.
Intelligent Interactions
Powered by Azure AI, CompaChat’s intelligent response generation adapts to your needs, offering smart solutions in real-time. Engage with AI that understands your audience and adapts to your needs.
Ready to revolutionize your communication Channels? Contact us for a demo or a POC and see CompaChat in action. Your journey towards result oriented communication starts here!
CompaChat
CompaChat Features
- Front end Connectors available for:
- WhatsApp Messenger
- Facebook Messenger
- Telegram
- SMS Gateway
- Mobile App
- Chatbot on your web page
- Microsoft Teams
- Others
- Comprehensive Role Based Security, allowing you to secure row-level data. Only authorized users will be able to see this app within their environment.
- RAG architecture for increased accuracy.
- Does not require Microsoft CoPilot.
- Your documents are not required to be stored in SharePoint or any other shared location.
- Your confidential information does not leave the four walls of your organization.
- Generates responses with actionable adaptive cards, so users can just click on a URL and take recommended action(s).
- CompaChat is a Caresoft Product. It comes with full support from your trusted and time-tested vendor
CompaChat Architecture
Front End
Middle Tier – It has the following Components –
- Pre-Processor – It scans prompts for adversarial attack attempts, applies role based security, conducts Responsible AI Checkup and logs incoming messages.
- RAG Enhancer – Using Azure AI Search, it performs Vector Search on pre-indexed internal data to prepare Grounding Data.
- Response Generator – Using Azure Open AI LLM, it generates Natural Language Response.
- Post Processor – It conducts Responsible AI Checkup on the response generated by CompaChat, generates additional content with actionable links on adaptive cards, applies customer branding, logs the output, and prepares the output for the Front End.
Back End – It has the following Components –
- CompaChat Search Engine – It uses Azure AI Search to perform Vector Search on pre-Indexed Data.
- CompaChat OpenAI LLM – It utilizes Azure OpenAI LLM to prepare Natural Language response based upon Users prompt, Grounding Data, System Messages and the context.
- CompaChat Vector Indexes – They store your internal data relevant to the application. The indexes are pre-populated from your internal databases, documents, Websites, as well as external data sources.
Administration
- CompaChat Vector Indexer – It creates embeddings from internal documents (pdf’s, websites, structured and unstructured data sources) and pulls relevant data from various internal and external sources to create Vector indexes. The indexes are refreshed at a pre-determined interval.
- CompaChat Admin Dashboard – Configuration, Administration and Analytics – Usage, Responsible AI and Security Data
- CompaChat uses private instances of Azure AI Search and Azure Open AI Services for LLM Chat features. Thes instances reside and run within your organization’s Microsoft Tenant.
- Your Confidential Documents and relevant data from your internal databases are vectorized and the embeddings are stored and indexed in an Azure Blob storage within your organization’s four walls.
How it works
A User starts a conversation using their favorite channel, such as Facebook Messenger, WhatsApp, Telegram, SMS, Mobile App, Chatbot on your web page or MS Teams.
The input message is passed through the CompaChat Connector for User’s Channel to the CompaChat Receiver in your environment.
CompaChat checks user’s credentials and validates their access using role based security. It validates user input to neutralize adversarial intents and checks input prompts applying Responsible AI principles and logs input prompts (optional).
CompaChat searches your internal documents / internal database based upon security level assigned to the user for relevant information and prepares Grounding Data. (CompaChat supports row-level security).
CompaChat modifies user’s input prompt with grounding data and appends system messages as well as the context. It passes the modified prompt to CompaChat Open AI LLM which uses a private instance of Azure Open AI service. CompaChat Open AI LLM prepares a Natural Language Response.
CompaChat checks the response against Responsible AI principles, generates additional content with actionable links on adaptive cards, formats the response using customer’s branding requirements, logs the output (optional) and returns it to the front end with actionable adaptive cards. The User can continue the conversation or take recommended action(s) by clicking on the links in actionable adaptive cards.
CompaChat allows the user to carry on their conversation by forwarding their previous interactions from the session as additional context during the subsequent calls.
Use Cases
Internal Communication
Retail
- Connect with your shoppers using your mobile Apps, Web page, FB Messenger, WhatsApp, SMS, Telegram or any other channel.
- Replace the search button in your webstore with a customer friendly chatbot to assist shoppers.
- Increase sales with consultative selling. Ex. “Recommend me a product for my joint pain”.