Bring Your Own Orchestrator with Teams AI Toolkit #13973
-
|
Hi all, Has anyone come across a concrete example of implementing a "bring your own orchestration" pattern with the Teams AI Toolkit? This option is often mentioned in the official presentations and slide decks, but I haven’t seen a clear implementation. Our use case is to have the toolkit delegate the user's input to a custom orchestrator, which would handle agent coordination and return a response back to the toolkit for display. If anyone has implemented this pattern or can point to a sample or guidance, it would be much appreciated. Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
This means you would not use the AI features of Teams AI Library. Instead, you would use the Application object and adapter to handle the conversation / messages only and use those as inputs with something else, like Semantic Kernel. The other option is to keep the Teams AI planner system and use it with a custom planner that uses another toolchain, like LangChain. There is an example of this pattern here: https://github.com/microsoft/teams-ai/tree/main/python/samples/07.planners.a.langchain |
Beta Was this translation helpful? Give feedback.
This means you would not use the AI features of Teams AI Library. Instead, you would use the Application object and adapter to handle the conversation / messages only and use those as inputs with something else, like Semantic Kernel.
The other option is to keep the Teams AI planner system and use it with a custom planner that uses another toolchain, like LangChain. There is an example of this pattern here: https://github.com/microsoft/teams-ai/tree/main/python/samples/07.planners.a.langchain