Search through more than a hundred articles on every aspect of

Greg Warzecha
Written by Greg Warzecha

AI Chat introduction

Today, I am thrilled to unveil our newest product, a revolutionary innovation that will change the way you think about Chat. 

It is sleek, intuitive, and user-friendly, with cutting-edge technology that sets it apart from anything else on the market. We've spent countless hours perfecting every aspect of this product, and I am confident that it will exceed your expectations. It is a true testament to our commitment to excellence and our passion for creating products that improve people's lives. 

AI Chat utilizes advanced deep learning LLM (Large Language Models) to provide a seamless and personalized support experience for users.

System is powered by two models:

  • first, an OpenAI Davinci-003 model trained for,
  • second, the Angel I model, which is proprietary and hosted on our company's servers in Europe.

The capabilities of our AI Chat include:

  • automatic reading and understanding of user messages, and the generation of appropriate responses based on the user's message and the knowledge base available within the application,
  • the ability to easily add new knowledge to the system by creating articles on specific topics. For example, if we wish for our AI Chat to respond to questions regarding a new "Orders" module, we can simply create an article containing the relevant information and the system will use this knowledge to generate responses to user questions,
  • AI Chat is currently able to respond to questions in English, with plans to expand to other languages in the near future.

To use our AI Chat, users must first enable the feature for their application. Once the module is available to all users, we will also provide the option to enable it from the control panel.

Users have the option to enable the AI Chat for "selected conversations" or "all conversations". The option to enable the AI Chat for "selected conversations" will appear in the input bar in the Inbox.

Frequently Asked Questions:

Does the system learn from past conversations?

The system currently only learns from the knowledge base.

Can I specify how the AI Chat should greet or farewell customers?

Yes, you can create an article outlining the guidelines and procedures for customer support, and the AI Chat will follow these rules when responding to customers.

Will the AI Chat be able to respond to questions about orders?

This capability will be available in version 2.0.

Is there a risk of the AI Chat responding with inappropriate content?

Currently, we use a pre-trained model from OpenAI, which minimizes this risk. However, if we use our proprietary model in the future, there may be a slight chance of such occurrences.

Does the AI Chat respond when an agent replies?

No, the AI Chat only generates a response when a user initiates the conversation, with an expected response time of 2-5 seconds.

What if there is no response from the AI Chat?

This means that the model is unable to understand the user's message or does not have enough knowledge to generate a response.

Can I use the AI Chat for automating responses to customers?

This capability will be available in version 1.5.

Thank you for joining us on this journey, and I can't wait to see the impact this product will have on your company.