Use Joule Studio to Create an Assistant to Monitor Supplier Risk

1. Open Joule Work and select the Develop area.

   

2. On the Agent tile, choose Create.

3. Keep New Solution in the Solution dropdown, and fill in the agent details.

   Name:

   COPY Copy

   Supplier Performance Assistant

   Intent Statement:

   COPY Copy

   Create an AI agent that monitors supplier performance, identifies at-risk suppliers, and recommends approved alternative suppliers using simulated procurement data.

   Select the Quick Create checkbox below the intent statement. This tells the tool to skip clarifying questions and move through all phases automatically.

   

4. Choose Create.

1. Open Joule Work and select the Develop + area.

   

2. On the Agent tile, choose Create.

3. Keep New Solution in the Solution dropdown, and fill in the agent details.

   Name:

   COPY Copy

   Supplier Performance Assistant

   Intent Statement:

   COPY Copy

   Create an AI agent that monitors supplier performance, identifies at-risk suppliers, 
   and recommends approved alternative suppliers using simulated procurement data.

   Leave the Quick Create checkbox unchecked. This tells the tool to ask you clarifying questions before generating the intent, and to pause for your confirmation at each phase.

   

4. Choose Create.

This is where the tool tries to understand your intent and map the challenge to SAP’s Reference Business Architecture. It has access to SAP Knowledge Graph, SAP LeanIX, and SAP Domain Models to help it create the intent document. Intent fit indicates how closely the proposed solution corresponds to your requirement.

In Quick Create mode, the tool skips the clarifying questions and generates the intent document directly.

This is where the tool tries to understand your intent and map the challenge to SAP’s Reference Business Architecture. It has access to SAP Knowledge Graph, SAP LeanIX, and SAP Domain Models to help it create the intent document. Intent fit indicates how closely the proposed solution corresponds to your requirement.

The tool asks you a few clarifying questions generated from your intent statement (the exact questions vary each run). Answer them using your judgement - the more complex your scenario, the longer generation takes.

Once you’ve answered the questions, the tool generates the intent document with a summary of who it’s for, what it does, how it’s built, and where it fills gaps in the SAP standard offering.

Review the intent document. You can explore it in detail on the Idea Board on the Intent tab.

If anything doesn’t reflect your goal accurately, refine it by continuing the conversation in the chat. Once you’re satisfied, proceed to the requirements phase.

After the intent is ready, the tool automatically moves on to requirements generation.

While the requirements are being generated, you can explore the intent on the Idea Board on the Intent tab.

When the requirements are ready, you have the opportunity to review and refine the Product Requirements Document (PRD). For this tutorial, accept it as-is and move on.

At this stage, you can see your PRD in markdown format similar to the one below. If you need to update it manually, you can click on the text in the view or edit the file in the dedicated Code tab.

After the intent is ready, enter the following in the chat when the tool prompts you:

Create requirements

When the requirements are ready, review the Product Requirements Document (PRD). If something is missing or incorrect, you can update it directly by clicking the text in the view or editing the file in the Code tab. Once you’re happy with it, continue to the specification phase.

After the PRD is ready, the tool automatically moves on to the specification phase.

While the specification is being generated, you can explore it in the Code tab. You’ll find it as specification/specification.md.

After the PRD is ready, enter the following in the chat when the tool prompts you:

Create specification

When the specification is ready, review it in the Code tab - you’ll find it as specification/specification.md. If you’d like to adjust any part of it, edit the file directly before moving on. Once you’re satisfied, proceed to building the solution.

In Quick Create mode, the tool automatically starts building the agent for you.

The tool now builds the solution - setting up the agent, creating the data and tools, and writing and running the tests. On the right, you can follow along. The exact steps and file names vary each run.

When all tasks are complete, you can review the generated solution. For this switch to the Solution step.

Explore the generated code in the Code tab to see how the agent is structured.

Choose the Try tab at the top to interact with the agent in a chat box.

Enter a question to test the agent. For this tutorial, try:

review all suppliers and their KPI scores

The agent processes your request and responds. What it can do depends on what was implemented - the right-hand panel shows a summary of the agent’s capabilities and example questions you can ask. The Traces panel at the bottom shows how the agent handled each request.

When the specification is complete, you could pass it on to another team to do the implementation. In this tutorial, you’ll get the tool to build the agent for you - choose the Build button in the top-right corner, or enter Build the solution in the chat when the tool prompts you.

The tool now builds the solution - setting up the agent, creating the data and tools, and writing the tests. On the right, you can follow along. The exact steps and file names vary each run.

When all tasks are complete, you can review the generated solution.

Explore the generated code in the Code tab to see how the agent is structured.

Once the solution is ready, choose the Try tab at the top to interact with the agent in a chat box.

Enter a question to test the agent. For this tutorial, try:

review all suppliers and their KPI scores

The agent processes your request and responds. What it can do depends on what was implemented - the right-hand panel shows a summary of the agent’s capabilities and example questions you can ask. The Traces panel at the bottom shows how the agent handled each request.