Learn Oracle

June 14, 2025

Introduction

AI isn’t just a buzzword anymore - it’s reshaping the way we work. With the rise of generative AI, especially through AI agents, businesses are now looking beyond traditional automation. But let’s be real: the challenge has always been about making these tools work within the systems we already rely on - securely, efficiently, and without extra cost or complexity.

That’s where Oracle’s AI Agent Studio, built natively into the Oracle Fusion Cloud suite, stands out. If you're already running Oracle Fusion Apps - be it ERP, HCM, SCM, or CX - this new capability offers a smart, scalable way to bring AI agents into your daily operations. And no, you don't need to bolt on third-party tools or invest in massive infrastructure overhauls.

Why AI Agents Are a Game-Changer

Unlike the static, one-size-fits-all automation tools of the past, AI agents powered by large language models (LLMs) like OpenAI or Llama are more intelligent and context-aware. Here’s what sets them apart:

They’re goal-driven – Think of them as colleagues who understand the “why” behind the task.
They work autonomously – Agents can act, decide, and even escalate or collaborate.
They specialize – Whether it’s HR, finance, or customer support, agents can be designed for specific personas.
They’re conversational and contextual – They engage in human-like conversations, remember prior interactions, and tailor responses using real-time enterprise data.

This makes them ideal partners in managing not just tasks—but end-to-end business processes.

What Makes Oracle AI Agent Studio Different

What struck me most about Oracle AI Agent Studio is that it’s not some add-on. It’s embedded into the Fusion apps you’re already using. Here’s what you get out-of-the-box:

Deep Fusion integration: Agents use the same business roles, workflows, and objects already in place. No custom integration nightmares.
Zero extra infrastructure: Everything runs within your existing Fusion environment, including the LLMs.
Speedy deployment with prebuilt templates: From single-agent tasks to multi-agent teams, you can roll things out fast.
Fully customizable: You can tweak Oracle-provided agents or build new ones from scratch—change prompts, insert approval checkpoints, or redefine workflows.
LLM flexibility: Use Oracle-optimized models or plug in others that suit your domain.
Rich toolset: Whether it’s pulling business data, accessing documents, emailing users, doing calculations, or calling external REST APIs - it’s all available.
Built-in governance and compliance: Agents respect Fusion’s existing access controls, roles, and security configurations.
Robust testing tools: Simulation environments and dashboards let you test, monitor, and refine agent behavior before going live.

Use Cases That Actually Matter

What’s exciting here isn’t just the tech - it’s what you can do with it. A few real-world examples:

📌 In Human Capital Management (HCM)

Imagine an employee trying to understand benefit options after a major life event. An AI agent steps in, fetches personalized info, explains options in plain language, and even remembers the context for future follow-up. Agents can also support tasks like updating personal details, recommending learning modules, or guiding through compensation changes.

⚙️ In Supply Chain Management (SCM)

A Technician reports a faulty machine. An agent takes the description, diagnoses the issue using manuals and historical data, initiates a work order, and notifies procurement for part replacement. It’s end-to-end resolution without switching screens.

Or a customer wants to understand the return options on the orders placed.

💰 In Enterprise Resource Planning (ERP)

An AP agent automatically processes invoices—matching lines, predicting GL codes, or escalating exceptions for approval. Something that usually takes days now wraps up in hours.

💼 In Customer Experience (CX)

Connected equipment goes down? An AI agent can run diagnostics, order parts, create a service request, and even notify the customer. And if something goes off-script - say a part isn’t available - it loops in a human rep with recommended actions.

These are a few handful examples. Oracle does provide many such tailored templates out of box.

What This Means for Different Roles

The impact of Oracle Fusion AI Agents extends across the organization:

End users enjoy simplified, smart, and responsive experiences.
Managers get the flexibility to quickly adjust processes to changing business needs.
Developers finally have tools that allow them to build meaningful AI features with real enterprise context.
Leaders gain measurable ROI from automation that actually works—without compromising on security or compliance.

Final Thoughts

AI agents aren’t just about replacing routine tasks - they’re about reimagining how we work. With Oracle AI Agent Studio, you don’t have to reinvent your tech stack. You just unlock more power from the one you already have.

If you're already using Oracle Fusion Cloud, this is an opportunity to move from automation to intelligent collaboration - putting AI agents to work with your team, not just for them.

And the best part? You can start now - without waiting on a giant digital transformation roadmap.

May 24, 2025

Here's a step-by-step guide for POC in Oracle Cloud SCM, focusing on creating a new item, creating a sales order, and shipping it.

Phase 1: Setup and Item Creation

Verify Initial Setup (Pre-requisites):

Access: Ensure you have the necessary roles and access to Oracle Cloud SCM (e.g., Inventory Manager, Order Manager, Product Data).
Business Unit (BU): Confirm you have a Business Unit defined and associated with your inventory organization.
Inventory Organization: Ensure your target warehouse is set up as an Inventory Organization in Cloud SCM. You mentioned leveraging an existing one, so just confirm its existence and setup.
Item Master Organization: Identify your Item Master Organization. This is where you create the item definition centrally.
UOMs (Units of Measure): Basic UOMs (e.g., EA for Each, LB for Pound) should be set up.
Item Catalogs/Categories: While not strictly mandatory for a basic POC, it's good practice to have some item catalogs and categories defined.

Create New Item:

Navigation: Go to Product Management -> Product Information Management.
Create Item:

Click on "Tasks" panel (usually on the right or left) and select Create Item.
Organization: Select your Item Master Organization.
Item Class: Choose an appropriate Item Class (e.g., Purchased Item, Finished Good). This drives default attributes and lifecycle.
Template (Optional but Recommended): If you have a template for similar items, use it to pre-populate attributes.
Item Number: Enter a unique item number.
Description: Provide a meaningful description.
Lifecycle Phase: Set to New or Active.
Status: Set to Active.
Primary Unit of Measure: Select the base UOM (e.g., EA for Each).
Save: Save the item.

Associate to Inventory Organization:

After saving, go to the Associations tab for the item.
Click the + icon to add an association.
Select your Inventory Organization (the warehouse where you want to transact this item).
Enable for Inventory: Ensure the Inventory Asset Value (or similar) attribute is checked at the organization level if you want it to be an inventory item.
Order Management Enabled: Crucially, for Order Management, navigate to the Sales and Order Management tab for the item (at the org level) and ensure Customer Ordered, Customer Orders Enabled, Internal Ordered, and Internal Orders Enabled are set to Yes.
Pricing Enabled: On the Sales and Order Management tab, ensure Priced is set to Yes.
Save: Save the association.

Define Inventory Attributes:

Go to the Specifications tab -> Inventory section.
Review and adjust attributes like Min/Max Quantity, Lot/Serial Control (if applicable), Picking UOM, Storage UOM. For a POC, defaults might suffice.

Define Sales & Order Management Attributes:

Go to the Specifications tab -> Sales and Order Management section.
Confirm the Shippable, Customer Orders Enabled, Customer Ordered attributes are set to Yes.
Ensure a Default Sales UOM is set.

Costing (Optional for basic POC):

Go to the Costs tab. You can manually enter a standard cost or wait for costing processes to run. For a basic order, it's not strictly necessary to have a full cost rollup, but it's good practice.

Phase 2: On-Hand Quantity & Pricing

Load On-Hand Quantity (Simulate Receipt):

Method: Miscellaneous Receipt:

Navigation: Supply Chain Execution -> Inventory Management.
Tasks: Create Miscellaneous Transaction.
Transaction Type: Miscellaneous Receipt.
Item: Select your new item.
Quantity: Enter the desired on-hand quantity.
Subinventory: Select a valid subinventory.
Locator (if applicable): Select a locator.
Reason Code: Select a reason.
Submit.

Verification: Confirm on-hand quantity by searching for the item in Manage Item Quantities.

Set Up Item Price:

Navigation: Order Management -> Pricing Administration.
Price List: Identify an existing Price List or create a new one for your POC.

Go to Manage Price Lists.
Select or create a Price List (e.g., "Corporate Price List").

Add Item to Price List:

Edit the chosen Price List.
In the Items tab, click + to add a new price list line.
Item: Select your new item.
UOM: Select the selling UOM.
Price: Enter the selling price for your item.
Start Date/End Date (Optional): Set effective dates if needed.
Save and Publish: Save your changes and ensure the Price List is Published.

Phase 3: Order to Ship Process

Create Sales Order:

Navigation: Order Management -> Order Management .
Create Order: Click on Create Order.
Header Information:

Customer: Select your existing customer.
Sold-to Business Unit: Select the appropriate BU.
Bill-to Account/Ship-to Account: These will likely auto-populate based on the customer.
Requested Ship Date: Enter a date.
Price List: Ensure the correct Price List (with your item's price) is selected.

Line Information:

Go to the Lines tab.
Click + to add a line.
Product: Search for and select your new item.
Quantity: Enter the desired order quantity.
UOM: Select the selling UOM.
The price should auto-populate from your price list.
Shipping Details: Ensure the Ship-from Warehouse (your inventory organization) is correctly defaulted or selected on the line.

Save and Submit:

Click Save to save the draft order.
Click Submit to send the order through the order orchestration process.

Monitor Order Status: Observe the Fulfillment Lines status. It should progress from Awaiting Scheduling to Scheduled, then to Awaiting Shipping.

Pick Release (Generate Pick Slips):

Navigation: Supply Chain Execution -> Inventory Management.
Tasks: Manage Shipments or Pick Waves.
Create Pick Wave:

You can create a new pick wave.
Select Release Sales Orders.
Enter criteria to find your order (e.g., Order Number, Warehouse, Customer).
Specify Release Rules (e.g., Auto Confirm Picks for simplicity in POC).
Click Release.
Alternatively, the order might be released automatically if you have auto-release rules configured.

Confirm Pick Slip:

After pick release, the order line status should move to Awaiting Shipping.
If Auto Confirm Picks wasn't used, you'd typically confirm picks in Confirm Pick Slips or Review Completed Picks. For a POC, if you can skip physical picking and confirm directly, that's quicker.

Ship Confirm:

Navigation: Supply Chain Execution -> Inventory Management.
Tasks: Manage Shipments.
Find Shipment: Search for your shipment using the Order Number or Shipment Number (which is generated after pick release).
Edit Shipment: Open the shipment.
Lines Tab: Review the items and quantities.
Ship Confirm: Click on Ship Confirm.

This action will typically update inventory, generate packing slips, and trigger invoicing interfaces.
Ensure Actual Ship Date and Shipment Method are populated.

Verification:

Check the order status in Order Management; it should now be Shipped and the Interfaced.
Verify the on-hand quantity for your item in Inventory Management has decreased.