Use an AI Agent to Help Create Work Instructions for an Operation

Welcome to the demo of the 25B feature Use an AI agent to help create work instructions for an operation in Fusion Manufacturing. With this feature, you can use an AI agent to help create work instructions for an operation. The AI agent can specialize in company-specific policies and procedures, which can help manufacturing engineers create accurate work instructions for operators.

Manufacturing engineers rely on following procedures and compliance documentation, and most often end up searching through large, varied sets of documents before they can determine the correct steps to perform in an operation. Now, the users can access AI agent on manufacturing pages through a page-level guided journey. The agent is conversational and answers questions from the knowledge base.

The AI agent also includes citations from the source documents. First, we will see how the AI agent can help create work instructions for an operation. I logged in as a manufacturing engineer. I launched the work definition work area under Supply Chain Execution and then launched the task Manufacturing Work Definitions.

Now, let us open the work definition for the product MR-Laptop and add some instructions with some help from the AI agent. This laptop has only one operation. I will create some instructions for operation 10. To create work instructions, I go to the Instructions tab, select an operation, and click Add Instruction button. I can access the agent through the guided journey provided in this page.

Since this is an assembly operation. I use the Agent to find out the steps required to start the assembly process. The agent reads through the procedures and informs that case preparation is the first step in the assembly. Now, I want to know the details of that step, which I simply ask the agent to provide.

The agent shares the instructions to carry out this step. I can simply copy this information and put that as instruction text for my work instruction. I can choose to expand the sources for more detail. Being able to extract readily usable text throughout the authoring process helps save time in creating the entire set of work instructions. This concludes the demo of how we use agents in authoring of instructions.

Let us see how we can set up the agents. This requires the following steps. Defining tools and adding documents to the tool, publishing the documents, mapping tools to the AI agent. To set up intelligent agents, I will launch this action under My Client Groups Configure HCM agents. This requires the privilege manage intelligent agent. First, I will go to the Tools tab using the navigation at the bottom of the page and add a new tool.

I will provide the name and description details of the tool. Then add a document to the tool and start attaching files to the document. After saving the attachments in the document, I can click the Add button to save the tool. I open the tool I just created and changed the status of my documents to ready to publish to prepare for publishing these documents.

To publish these documents, I go to Schedule process page under Tools. I find the process agent documents process and run it and make sure it succeeds. I can go back and see in my tool that the process successfully published my documents. Now, I can switch to the agent tab and add a new agent.

I have provided the name, code, and description. For this agent, I am going to name it and map the tool we previously defined. You can add more than one tool to an agent. Click Add button on top to save the agent. The final step, we will do is to deploy this agent into a guided journey for a page. To do that, we will create a guided journey, add an agent task to the journey, deploy the journey to an application page.

Launch guided journey task for my Client Groups. Click Create to create a new journey. I created a journey by providing name and code details. For this journey, let's add a new task. Provide the task details. I have provided the name, code, and description. Select task type as agent.

This task type is enabled through agent task type enabled profile option. After choosing agent task type for configuration, pick the tool we just created and published. After saving the tasks, activate this journey. You can have more than one agent task in a journey or a combination of agent and other task types.

If your journey code matches default journey code for a page, it should appear automatically in the application. Otherwise, you can launch the DB editor express mode as an admin. Here, you provide the journey code and publish it for the users. Users need REST service-guided journey read-only privilege to access the journey and access intelligent agent chat privilege to access the agent.

Here is a quick recap of all the setup steps that we just went through. To define the agents, we launched manage intelligent agents task to create a tool and added our documents there. We ran an ESS job to publish the documents. We created an agent and mapped our tool to it.

To deploy the agent to the manufacturing page, we created a guided journey. We added a task of the type agent. We deployed this journey to a page using Visual Builder. This concludes the demonstration for this feature. Thank you for watching.