Generative AI

Generative AI Integration

Augmenting the user experience on conversations across multiple platforms (WhatsApp, Web, etc.) by generating contextual responses on top of the conversational AI-based responses. Enhancing performance by allowing chatbot to summarize, explain and generate an appropriate response. You can train chatbots with content from the website, documents, or product catalogs. Enabling the chatbot to respond to the user queries across multiple channels improving the engagement and effectiveness of conversions.

ChatGPT Integration

To initiate the integration of ChatGPT, begin by navigating to the Orai portal. Once there, access the “Integration” section to establish a connection for ChatGPT integration. In this step, you will be prompted to add an account specifically dedicated to the integration process.

To proceed with the ChatGPT integration, locate and click on the “GPT – Connect” button within the Orai portal. This action will prompt you to add the necessary account details for the integration process.

Choose the integration type by selecting between “OpenAI” or “Azure OpenAI.” This step involves specifying the platform for which you are configuring the integration within the Orai portal.

When selecting OpenAI integration, you will be required to input either the account name or the API key.

Once you have entered and verified the required information, your account will be successfully added to the GPT integration. This ensures that Orai can now effectively utilize the specified settings for GPT functionalities.

Once the GPT is integrated, the document to train the bot can be uploaded.
To upload a Generative AI document, navigate to the training section and choose the “Generative AI” option. This selection allows you to upload and incorporate documents that utilize Generative AI capabilities during the training process.

Select the “Generative AI” option, and within this section, you will find various features including option settings, document categories management, and document uploading. Explore these functionalities to customize your Generative AI experience.

In the settings, you can customize fundamental details such as the domain type (e.g., eCommerce, finance, healthcare), assistant role, LLM (Large Language Model) provider name, chat model name, embedding model, response format (e.g., paragraph, bullet points), and response length (e.g., concise, detailed). These options allow you to tailor the Generative AI to your specific preferences and requirements.

Document Upload

In the Docusense upload section, you have the flexibility to choose the document uploading method. You can opt for a single upload using a URL or file, or you can perform multiple uploads. To upload a document, enter its name, select the category, specify the language (if applicable), and then proceed to upload the document either through a file or URL format. It’s important to note that documents should be in PDF format. Once the details are provided, click the save button to complete the upload process.
If there are errors in the document, an error will be prompted. Otherwise, once conditions are met, the document will be successfully uploaded and ready for search functionality.

 

Once the bot is updated it will start giving the responses from the trained document.

Getting the OpenAI API

You can find comprehensive information about the API on the OpenAI website at both https://openai.com and https://platform.openai.com/docs/introduction . For details about the API itself, visit https://openai.com/product , and for pricing information, head to https://openai.com/pricing .