Pro-Code vs Low-Code Comparison - Multi-agent decision support

Summary

Choosing between a pro-code and a low-code approach is a decision teams make over and over, and it drives development speed, total cost of ownership, scalability, governance, and the skills a team needs to hire for. The answer usually depends on who you ask.

This sample settles that with a multi-agent pattern in Microsoft Copilot Studio. An orchestrator agent takes the user's question, triggers a pro-code child agent and a low-code child agent in parallel with the same normalized prompt, and merges their two perspectives into a single structured report.

Every answer follows the same shape, so two different questions produce reports you can put side by side:

The screenshots come from the original recording of this scenario, where the orchestrator still carried its development name, Comparing Agent Production. It is named Pro-Code vs Low-Code Comparison in this sample.

Contributors

• Keshav Keshari

Version history

Prerequisites

• A Microsoft Copilot Studio license
• A Power Platform environment with Dataverse
• Power Platform CLI to import the solution

Minimal path to awesome

• Clone this repository (or download this solution as a .ZIP file then unzip it)

Copilot Studio using Solution Import

This sample uses the Power Platform CLI to import the agent. For installation instructions see What is Microsoft Power Platform CLI?

• Ensure you are authenticated with pac auth create

cd samples/mcs-procode-vs-lowcode-comparison

# Package up the solution. Pointing at ./src is important.
pac solution pack --zipfile mcs-procode-vs-lowcode-comparison.zip --folder ./src

# Import into a specific environment with -env, or leave it off for the default environment
pac env list
pac solution import --path ./mcs-procode-vs-lowcode-comparison.zip

• Open Copilot Studio, select the environment you imported into, and open the Pro-Code vs Low-Code Comparison agent
• Select Publish to make the agent available
• Open the Test your agent pane and try one of the conversation starters, or ask your own comparison question

Features

Using this sample you can build a Copilot Studio agent that:

• Coordinates two child agents from a single orchestrator, rather than packing every perspective into one set of instructions
• Triggers both child agents in parallel with the same normalized question, so neither perspective sees the other's answer first
• Keeps each perspective's expertise scoped to its own agent, which keeps the instructions short and focused
• Returns a consistent, tabular report for every question, so answers are comparable over time
• Names the losing option's specific constraints and the levers that would change the verdict, instead of stopping at a recommendation

Agents

Each child agent is stored as a botcomponent of type 9 with kind: AgentDialog, alongside the orchestrator's gpt.default component and the system topics.

Solution structure

Help

We do not support samples, but this community is always willing to help, and we want to improve these samples. We use GitHub to track issues, which makes it easy for community members to volunteer their time and help resolve issues.

You can try looking at issues related to this sample to see if anybody else is having the same issues.

If you encounter any issues using this sample, create a new issue.

Finally, if you have an idea for improvement, make a suggestion.

Disclaimer

THIS CODE IS PROVIDED AS IS WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING ANY IMPLIED WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR NON-INFRINGEMENT.