Chatbots for customer support and service
TEXT_INPUT:Ask the user to provide their name.;;CHATGPT:Prompt ChatGPT to greet the user and ask how it can assist them. Reference the user's name from the previous step.;;HTTP_REQUEST:Send a request to the customer support API to retrieve relevant information or FAQs based on the user's inquiry.;;CHATGPT:Prompt ChatGPT to provide the user with the retrieved information and ask if they need further assistance.;;DISPLAY_OUTPUT:Display the responses from ChatGPT to the user in a user-friendly chat-like format.
Generative AI models like Language Model (LMs) and image generators can be used in Chatbots to enhance customer service and support. For example, ChatGPT, powered by OpenAI, can be configured for customer support Chatbots, offering rich and comprehension responses to customer queries. This improves the customer support experience, by providing high-level natural language understanding and generating text that can match customer inquiries with relevant responses, FAQs, or guides.
In addition to text, Image generating AI models can aid in visual customer assistance. For instance, when a customer asks for guidance on setting up a product, the Chatbot can generate step-by-step instruction images to assist visually. These models can create custom images as per user queries, making problem-solving interactive and easy to understand.
Overall, these generative AI models can help in providing 24/7, personalized and efficient customer services and support.
How to build with Clevis
This is an illustrative example of an application you can build using Clevis - a powerful AI tool. The application is a customer support chatbot that operates through a series of steps to assist users with their inquiries. Applications like these in the automation or customer service area can also be created using Clevis.
Initially, the application functions by asking users to input their names. Next, the critically-acclaimed ChatGPT is activated to warmly greet the user while asking them how it can be of help. Importantly, the interaction at this step is personalized by mentioning the user's name obtained from the initial step.
Following the chat, an HTTP request is dispatched to a customer support API. The purpose of this request is to gather any relevant information or frequently asked questions that are aligned with the user's query.
Once the required data has been fetched, ChatGPT is again prompted. At this time, it provides the user with the obtained information, also asking whether they require additional support.
Finally, the chatbot displays the generated responses in a format that resembles a user-friendly chat system. This is designed to make it easier for users to track and comprehend the conversation.