Prompt Flow (Part 4)

Neil Haddley • August 24, 2024

Using Semantic Kernel's planner to automatically orchestrate multiple AI plugins so an agent can combine them to achieve a user's goal

AIprompt-flowplanningazure-aillmops

I used Semantic Kernel planning to combine multiple plugins so the AI agent can solve a user’s goal by orchestrating them together.

I selected the existing .venv environment

I created a selectedService variable

I created an ask variable (the goal)

I created a kernel

I added summarize, writer, text and shakespeare plugins

I created a Sequential PlannerI asked the planner to create a plan based on my ask (goal)

I printed out the planning stepsI invoked the plan and printed the result (a short poem in the style of Shakespeare translated to French and converted to all upper case)

I updated the ask (goal)

I added the fun plugin

I printed out the updated plan's steps

I invoked the plan and reviewed the result

I created a kernel

I created an (stub) Email plugin

I added the Email pluginI added the Math and Time plugins

I used the planner to answer three questions

Part of the Prompt Flow series

← Previous in series

Prompt Flow (Part 3)

Using Microsoft Semantic Kernel to build a plugin-based AI agent that generates jokes, called from an Azure Prompt Flow node

Next in series →

Prompt Flow (Part 5)

Creating and deploying an AI chat application with Azure Prompt Flow and LangChain in Azure AI Studio