Azure OpenAI Connected System

Overview

This plug-in provides Appian developers with direct access to Azure OpenAI’s services, including Chat Completions (ChatGPT), Completions, and Embeddings. Through this connected system, developers can enjoy all the features of Azure’s OpenAI services while also benefiting from the simplicity and ease of use that comes with Appian’s low-code platform.

Integrating these services into your Appian application enables a wide range of use cases.

• Chat Completions: The Chat Completions endpoint can be used for building chatbots, virtual assistants, text summarization, or any application that requires interactive conversations. An example using text summarization within an Appian app is available at the youtube link listed as a resource below.
• Completions: Integrating with the Completions endpoint will enable use cases such as drafting emails, generating code, writing articles, or any other text generation tasks.
• Embeddings: Lastly, the Embeddings endpoint facilitates various downstream tasks, including semantic search, text clustering, and recommendation systems.

Helpful resources:

• Setting up the Plug-In for Text Summarization of Record Data (Video Tutorial): https://www.youtube.com/watch?v=lQqFFM2T7-M
• Plug-In Documentation: available as a PDF through download in the App Market or through the Sample Application interface

Key Features & Functionality

• Chat Completions: This endpoint allows you to create completions using chat messages, utilizing ChatGPT and GPT-4 models. You can have interactive conversations by providing a series of messages as input to the model, which then generates a response.
• Completions: With this endpoint, you can provide a text command or prompt as input, and the model will generate one or more predicted completions based on the provided prompt. This can be useful for tasks such as generating text, drafting emails, writing code, and more.
• Embeddings: The Embeddings endpoint enables you to retrieve a vector representation of a given input. These vector representations, often referred to as embeddings, can be used by machine learning models and other algorithms for various downstream tasks such as semantic search, text clustering, recommendation systems, and more.