AI Agent: Planning Advisor for Exceptions

Supply Chain Planning Exceptions is one of the more important avenues for you as a planner to understand issues and opportunities with the current plan. However, the volume of exceptions hinders the efficient analysis of the plan. With this update, Planning Advisor for Exceptions and Notes, a new AI agent template is available to configure or use as-is for you to quickly understand the most important exceptions in the plan to be addressed. You can interact with the agent using natural language The template agent can be used to create an agent that will synthesize information and guide the users to summary insights by querying on the demand exceptions business object. The following additional business objects as applicable are available for you to extend the agent:

  • Supply exceptions
  • Resource exceptions
  • Supplier exceptions
  • Measure-based exceptions

This feature uses a Large Language Model (LLM) for responses to provide coherent responses to user queries through the user’s interaction with AI Agents. The AI Agents use the planning context such as Plan and Items to contextualize their responses to the user’s queries.

Planning Advisor – Use Plan Exceptions Assistant

In this update, the Planning AI advisor menu contains the Exceptions and Notes Assistants. The Exception assistant is seeded with an agent for Demand Exceptions. The Exception Agent can be accessed from the Planning Advisor menu after the configuration of the Agent team described in the ‘Setting up the AI Agent’ section.

Planning AI Advisor

Planning AI Advisor

You can query the agent with questions related to exceptions in your plan using human language sentences. Some sample questions:

  • What demands are at risk for Item MK-Eng-Computer ?
  • Give me all the exceptions in Org M1 for the month of January 2030 ?
  • Show late replenishment for sales orders for customer XYZ?
  • Show demand at risk due to a resource shortage on Jan 2030 in plan SP_Resources?
  • Show demand at risk due to insufficient lead time for item PPS-LDA-Computer and supplier Computer Peripherals Ltd. ?
  • What are Late Supply Pegged to Forecast before Jan 10 2030?

To improve the accuracy of the responses, the agent suggests clarifying questions for user’s additional inputs. For example,

Example query

Example query

Query response

Query response

In addition to the seeded Demand Exceptions agent, business objects for the supply, resource, supplier and measure based exceptions are also available in the business objects section of the AI Agent Studio. To know how to set up agents using these business objects, refer to Tips and considerations section of this document.

Setting up the AI Agent

Planning Advisor for Exceptions and Notes

The Supply Chain Planning Exceptions and Notes Advisor helps the users query and highlight potential issues related to exceptions generated by planning. It also helps planners query important planning assumptions and decisions documented as Notes for items in a plan.

NOTE: This template is common for Exceptions and Notes. Refer to the What’s New in AI AGENT: ANALYZE SUPPLY CHAIN PLANNING NOTES for more information on the usages of the AI agent for Notes.

Configuring the Planning Advisor for Exceptions

To use the AI Agent for Exceptions, follow the below steps:

Copy Template

Copy Template

  1. Make a Copy of the template and customize the name of your Agent Team by adding a suffix. This will be added to the template defaults names in templates properties to distinguish it from others. Duplicate suffixes are not allowed. This action navigates you to the Agent Teams page which shows the graphical representation of the Agent team.

Agent team

Agent team

Each of these agents has an assigned role:

  • SCP Exceptions Supervisor Agent: A Supervisor Agent is part of a multi-agent Supply Chain Planning system. It manages and orchestrates exception-related queries from a planning user by routing them to the correct Exception Worker Agent.
  • SCP Demand Exceptions Agent: A supply chain planning Demand exceptions agent interacts with supply chain planners to give them insights about their plan's or item's demand exceptions.
  • SCP Intent Classifier Agent: An Intent Classification Agent specializes in understanding exception-related queries in the Supply Chain Planning domain. It analyzes the planning user input and returns a structured representation of what kind of exception the user is asking about, and for which item(s) or plan.
  • SCP Notes Agent SCP: A Supply Chain Planning AI Worker agent which can retrieve notes information for given plan name or item name from planners.
  1. Publish the Agent Team: Click on the Publish action button. The agent team is now available for use.
  2. Copy the Agent Team Code. Click on the Agent Team’s Settings Icon on the page header. This opens the property details of the agent team. Copy the agent team code. This will be used for adding the Planning Advisor action to your page.

Copy agent team code

Copy agent team code

This opens the property details drawer for the Agent Team. This is required to copy into the planning page as Visual Basic constant property to enable the Agent in the planning menu interface.

  1. Update the SupplyPlanningAgentTeamCode parameter in the User Profile: To make the Exceptions agent available in the user menu, the agent team code needs to be added to the user profile. To achieve this, do the following actions:
  • Go to the user interface’s ‘Edit Page in Visual Builder Studio’ under the Administration setting.

User profile option

User profile option

This will initiate the Oracle Visual Builder Studio in a new browser tab with the below settings.

Using Visual Basic Studio to set a constant: There are now two possible ways that the VB Studio in the new browser tab could appear.

If the Visual Basic Studio project has already been set up in the environment, all you need to do is set the constant values (i.e., as seen in the screenshot below, set the Agent Team code to the correct plan type constant).

The Agent Team Code is the code that needs to be entered into that constant parameter. Update the SupplyPlanningAgentTeamCode with the agent team code obtained in step 3.

Update the SupplyPlanningAgentTeamCode parameter

Update the SupplyPlanningAgentTeamCode parameter

After doing this, publish the changes. You will now be able to see the Planning Advisor in the menu with the menu option for the Exceptions agent.

NOTE: If this is an environment where the VB Studio project has not been set up for “Edit Page”, you will be prompted to create new project, after which you will be taken to the VB studio setup page.

VB Studio setup message

VB Studio setup message

After the changes have been published, you will be able to see the Planning Advisor in the menu with the menu option for the Exceptions agent.

  • Reload the Redwood page. After reload, select the corresponding plan (of same type for which Agent Team Code was set)

Go to the menu at the top right -> then click on “Planning Advisor”

Planning AI Advisor menu

Planning AI Advisor menu

Accessing the Use Plan Exceptions Assistant initiates the agent interface. Clicking on “Use Plan Exceptions Assistant” or “Use Plan Notes Assistant” will open up the chat.

In the chat, you need not specify the plan name. As it has already been sent as context based on the currently open plan in the UIShell.

You are now ready to use the agent through a conversational interface.

 By allowing you to look into particular aspects of the exceptions, querying the demand exceptions agent accelerates the exceptions analysis process and helps you make crucial planning decisions faster, through a quick conversational experience.

Illustration of a user query and the Exception agent response

Illustration of a user query and the Exception agent response

Querying the exceptions agent

The queries are broadly classified into the following type of criteria patterns that can be used to ask the assistant questions:

  • Item criteria: By including item-specific questions in the query.
    • For example, What demands are at risk for Item MK-ENG_Computer ?

Query response:

What are the demands at risk for item

What are the demands at risk for item

  • Criteria other than Item: By including two other non-item specific criteria such as organization, customer, exception type or dates in the query.
    • For example, Give me all the exceptions in Org M1 for the month of January 2030 ?

Query response:

Exceptions for a period

Exceptions for a period

Queries that do not have specific criteria, prompt clarifying questions from the assistant, requesting more details to provide a response.

For example, Give me all the demand exceptions in the plan ?

This returns a message, "Can you please give more specific context like Item, Exception type, Organization, Date, Customer ?” to further refine the query.

NOTE: The results in the responses are not exhaustive and are prioritized by the top 5 results based on the order value of the exception result.

Steps to Enable

You don't need to do anything to enable this feature.

Tips And Considerations

You can use AI Agent Studio to use or copy a preconfigured agent template to create Agents for your business processes. To automatically add a suffix to all artifacts in your agent team, you can Copy Template instead of Use Template. When you copy a template, you're directly taken to the agent team canvas where you can edit the agent team settings, agents, tools, and topics. The Use Template option takes you through a step-by-step process for configuring each artifact in the agent team. For information on using AI Agent Studio, see How do I use AI Agent Studio?

Key Resources

  • Refer to What’s New in AI Agent: Analyze Supply Chain Planning Notes

Access Requirements

To access the Oracle AI Agent Studio for Fusion Applications and manage SCM AI agents, users must be assigned a configured job role that contains these duty roles:

  • SCM Intelligent Agent Management Duty (ORA_RCS_SCM_AI_AGENT_MANAGEMENT_DUTY)
  • Fai Genai Agent SCM Administrator Duty (ORA_DR_FAI_GENERATIVE_AI_AGENT_SCM_ADMINISTRATOR_DUTY)

To interact with AI agents in product pages, users must be assigned a configured job role that contains this duty role:

  • Fai Genai Agent Runtime Duty (ORA_DR_FAI_GENERATIVE_AI_AGENT_RUNTIME_DUTY)

To allow users to interact with agents, you must also enable permission groups in the Security Console on those users' configured job roles that contain the Fai Genai Agent Runtime Duty role. You can enable permission groups when you manage the basic information of your configured job roles.

 Users' configured job roles must also contain privileges that allow access to the pages where AI agents are enabled.