The AI capability 'suggested replies' in Deskhero works by leveraging OpenAI Embeddings API and a vector database. Each ticket resolution, knowledge base article, and scraped website data are sent to the OpenAI Embeddings API to generate embeddings vectors. These vectors are then stored in a vector database. When a query comes in, either through search or a new ticket, the query is run against the OpenAI Embeddings API to find similar relevant data. The relevant data is then sent to ChatGPT, along with the original question, to generate a response based on the attached data. This response is presented to the user as a search response or a suggested reply for tickets.
Use case 1
Enhanced Ticket Resolution - The AI capability 'suggested replies' enhances ticket resolution by analyzing previous ticket resolutions and generating suggested replies for support agents. This improves the efficiency and accuracy of ticket resolution.
Use case 2
Intelligent Search Results - When users search for information within Deskhero, the AI capability 'suggested replies' can provide human-like responses based on the relevant parts of the matched customer data. This helps users quickly find the information they need.
Use case 3
Continuous Improvement - By analyzing the data used for generating suggested replies, Deskhero can continuously improve its AI capabilities. The AI learns from ticket resolutions, knowledge base articles, and other data sources, allowing for ongoing enhancements and better performance.
