Copilot Studio: Add a Classify Node

Key insights

• Classify node: a built-in action in Copilot Studio that automatically labels unstructured text using an LLM.
  It replaces fragile keyword rules and routes messages to the right workflow steps.
• Copilot Studio Workflows: the environment where the Classify node runs natively, making AI a first-class workflow action.
  This reduces the need for external services and simplifies automation design.
• No custom training: you define categories with plain-language descriptions and optional examples instead of training a model.
  Adding examples improves accuracy but is not required to use the node.
• Semantic classification: the model classifies by meaning and intent, not by exact words or regex patterns.
  That helps correctly route varied requests like “my laptop won’t turn on” to a Hardware issue category.
• Setup and categories: add the Classify node to a workflow, choose the input text and AI model, then create named categories with clear descriptions.
  Use the dynamic content picker for inputs and add example texts to refine results.
• Routing and flexibility: the node shows one branch per category plus an Other (default) branch and supports many concurrent categories.
  You can attach multiple downstream steps to a branch, merge outputs later, and pick the model per node.

Keywords

add classify node Copilot Studio, Copilot Studio classify node tutorial, how to add a classification node to Copilot Studio workflow, create classify node in Copilot Studio, Copilot Studio workflow classify step, Copilot Studio AI classifier node, Copilot workflow classification tutorial, Copilot Studio node types