Microsoft Semantic Kernel is an open-source software development kit (SDK) that allows developers to build custom AI applications by combining large language models with conventional programming code and external data sources. It is a developer tool rather than an end-user product, used to build the kind of AI orchestration that powers more advanced custom AI agents and applications.
How Semantic Kernel works
Semantic Kernel acts as a bridge between an AI model, such as one provided through Azure OpenAI, and an application's existing code and data sources. It allows developers to define functions the AI can call, chain together multiple steps of reasoning and action, and integrate AI capabilities into existing software systems rather than building everything from scratch. This kind of orchestration is part of what makes more advanced AI agents possible.
How Semantic Kernel is used
- A software development team uses Semantic Kernel to build a custom internal application that combines AI-generated summaries with data pulled directly from existing business systems.
- A development partner builds a bespoke AI agent for a client using Semantic Kernel, orchestrating multiple steps including data retrieval, reasoning and action.
- A technology team uses Semantic Kernel to integrate AI capabilities into an existing line-of-business application without rebuilding the application from scratch.
- A business commissions a custom AI solution from a development partner who uses Semantic Kernel as part of the underlying technical architecture.
How Advantage approaches custom AI development
For most UK SMEs, low-code Power Platform tools such as Copilot Studio and AI Builder deliver the AI capabilities needed without requiring custom development. Where a genuinely bespoke AI solution is justified, Advantage works with development partners who use frameworks including Semantic Kernel to build it properly.
Frequently asked questions
Is Semantic Kernel a Microsoft product businesses buy directly?
No. Semantic Kernel is an open-source software development kit (SDK) that developers use to build custom AI applications. It is not a standalone product purchased by end users, but rather a tool used by development teams building AI solutions, often on top of Azure OpenAI.
How does Semantic Kernel relate to Microsoft Copilot?
Semantic Kernel is one of the underlying frameworks Microsoft uses to help build AI orchestration into its own products, including aspects of Copilot. Developers can also use Semantic Kernel independently to build their own custom AI applications with similar orchestration capabilities.
Do I need Semantic Kernel if I am just using Power Platform for AI?
No. Power Platform tools such as Copilot Studio and AI Builder are designed for low-code use without needing Semantic Kernel directly. Semantic Kernel is relevant for development teams building more bespoke, code-first AI applications beyond what low-code Power Platform tools provide.
