Software development companies can build their own agents with Semantic Kernel, an open-source SDK that empowers developers to create sophisticated agents within their apps. Agents are AI systems that can answer questions and automate processes for users. They range from simple chatbots to fully automated AI assistants.

Common scenarios to use Semantic Kernel

• You're building an agent that is a part of your own application (customer development tool or HR system), and you want people to stay in the realm of the same UI.
• You need full control on the orchestration engine, RAG implementation, model choices, and model parameters.
• With your agent service, you want to allow your customers to build extensions on top of your solution through plugins.
• Your solution utilizes canvases and other media than just text.

In this collection, you'll find a sample solution architecture, a fully working, cloud-ready application deployable with the GitHub Developer CLI to get you started, a fuller explanation of how to consider your scenarios and approaches, and some recommended Microsoft pages for a deeper dive into agents and plugins.

For a full explanation of all approaches and scenarios, you can visit the Creating Generative AI Experiences with the Microsoft Cloud. If you'd like help to streamline your app build, publish, and marketplace growth journey, visit app advisor on ISV Hub, answer a few questions, and get the right information you need, right now.