Why use the Generative AI block

Common issues in traditional chatbots 

1. Traditional bots may not interpret and respond accurately to complex queries.
   
2. The bot may not retain previous interactions, requiring users to repeat information.
   
3. When transferring to an agent, users may need to provide details again.
   

 
Key features   

1. Ensures smooth transitions between chatbot interactions.
   
2. Works across various platforms like WhatsApp and Web Chat while retaining context.
   
3. Can be added and set up within Guided Conversations without having to code.
   

Components of Zia Generative AI block

1. Input Variable Mapping: Input Variable Mapping allows you to pass user-provided or manually entered data to Zia for generating responses.
   
1. Prompt Field: Providing specific queries or instructions in the Prompt field helps Zia generate precise and relevant responses.
   
2. Context Field: Mapping the context will be used as knowledge that Zia will use to generate a meaningful response from.
   
2. Output Variable Mapping:
1. Helps in creating a variable for the response field and using it in a Display Message block to show the response provided by Zia.
   

Use the Zia Answer Bot if you want to answer customer queries from your KB content.

Use the Zia Generative AI block when you want more flexible, AI-powered responses based on user inputs and business context.

How to create a Generative AI block   

1. Go to Zia Generative AI block in the Integration section.
   
2. Add the get response from Generative AI block.
   
3. Give a suitable Block Name to identify.
   
4. Configure Input Variable Mapping
   
1. Map the Context Variable.
   
2. Enter specific instructions in the Prompt Field to help Zia generate precise and relevant responses.
   
5. Configure Output Variable Mapping
   
1. Create a Variable for the response field.
   
2. Use this variable in a Display Message block to show the response provided by Zia.
   

                                                                                                                        

How to set up the prompt field 

Tips for effective prompts

1. Be Specific: Use clear and detailed prompts to minimise ambiguity.
   
2. Include Variables: Leverage user inputs dynamically for personalized responses.
   
3. Avoid Open-Ended Queries: Keep prompts focused to help Zia generate precise answers.
   

1. Product Inquiry: "Tell me about the features of [Product Name]."
   
2. Troubleshooting: "How can I resolve error code 404 when accessing my account?"
   
3. Order Status: "Where is my order #[Order ID]?"
   

How to set up the context field 

Effective context configuration

1. Map Relevant Variables: Include past interactions, preferences, or any data fetched through APIs.
   
2. Limit to Essential Information: Focus on the most relevant details to avoid overwhelming the AI.
   

1. Company Info: "We offer a 30-day return policy."
   
2. Order History: "The user's last order was on [Date] for [Product]."
   

Transforming GC into an AI-Powered conversational bot 

1. Add a question block to gather user inputs effectively by framing questions.
   

1. Map input variables to use the data from the question block to configure the prompt field in Zia Generative AI.
   

1. Use a display message block to present the AI-generated responses to users clearly and concisely.
   

1. Add feedback options to use a Button Block to gather user feedback on the AI responses.
   

Implement a refinement loop

1. Add a jump block to enable users to refine their queries if they are dissatisfied with the initial response.
   
2. Redirect to the question block to map the refined input to Zia for an improved answer.
   

Best practices  

1. Ensure variables are specific and relevant for more accurate responses.
   
2. Provide only essential information in the context field to keep responses concise.
   
3. Use the Combine Block to set a detailed context with more characters or to include multiple variables in the context field.