Building Fusion Extensions with Redwood + AI Agents: A Smart Step Forward

 Oracle’s Redwood UX brought a new wave of design consistency and user delight to Fusion Applications. Now, with the rise of AI Agents in Oracle Cloud, we’re entering the next phase: smart extensions that are not just reactive, but proactive.

In this post, let’s explore how to combine Redwood-based Visual Builder apps with Fusion AI Agents to create intelligent, personalized, and impactful extensions — the kind that don’t just respond, but guide users.

Why Combine Redwood UX with AI Agents?

Modern enterprise users expect more than just pretty interfaces — they want smarter workflows, relevant insights, and less clicking. When you blend Redwood’s modern design system with AI Agents’ intelligence capabilities, you unlock:

• Dynamic recommendations

• Proactive nudges and alerts

• Intelligent data lookups

• Personalized guidance

• Faster decision-making

In short, you deliver extensions that think for the user, not just wait for input.

How It Works: The Smart Extension Model

Here’s the architecture at a high level:

1. Redwood App built in Oracle Visual Builder

2. Connects to Fusion’s REST APIs to read/write business data

3. Integrated with AI Agent Studio or prebuilt Oracle AI models

4. Optionally leverages Oracle Integration Cloud (OIC) for orchestrating multi-step flows

5. Embedded directly in Fusion (via Page Composer) or launched as a standalone app

This stack lets you deliver intelligent capabilities while staying fully within the Oracle ecosystem — no bolt-ons required.

Real-World Use Cases

1. Smart Project Costing Tracker

A Redwood app shows project spend vs. budget. AI Agent highlights overspending trends and suggests actions like freezing PO releases or reassigning tasks — all in real-time.

2. Guided Onboarding for Suppliers

AI Agent helps fill out forms, validates GST/VAT data, and nudges users to upload required documents based on country-specific rules. The Redwood UI ensures clarity and accessibility across roles.

3. Proactive Leave Planning Assistant (HCM)

Redwood extension suggests ideal leave windows based on team availability and auto-checks policy compliance. AI-driven logic helps reduce manual back-and-forth.

4. AI-Powered Invoice Exception Handler (Finance)

Invoices stuck due to mismatches are flagged by the agent. The Redwood app guides the AP user through resolution — pulling in supplier contact, prior approvals, and risk scoring.

How to Build One: Key Steps

1. Define the Use Case and Outcome

Focus on what you want to improve or predict. Is it faster decision-making, automated classification, or guided user actions?

2. Build the Redwood UI

Use Oracle Visual Builder:

• Redwood Web App template

• Connect Fusion data via REST APIs

• Ensure it’s responsive and user-role-aware

3. Integrate the AI Agent

• Use AI Agent Studio to design an agent (intent, rules, models)

• Connect it to relevant Fusion datasets

• Decide whether to embed the agent in the UI or trigger it behind-the-scenes

4. Connect the Dots

• Combine the Redwood UI with real-time outputs from the AI Agent

• Use toast messages, guided steps, or side panels to display agent feedback

• Optional: Use OIC for orchestration if multiple apps/systems are involved

5. Embed & Test

Deploy the Redwood app and embed it into Fusion (via Page Composer or Navigation link). Simulate real user scenarios to test:

• Agent accuracy

• UI responsiveness

• Business rule alignment

• Role-based visibility

Tips for Success

• Start small — one agent, one outcome, one page

• Work closely with business users to define realistic agent behaviors

• Use Redwood design tokens and layouts for consistency

• Monitor performance, especially when calling models or APIs in real-time

• Ensure proper error handling and fallback options

Summary

Combining Redwood UX with AI Agents moves us from automation to augmentation. You’re not just building tools — you’re building assistants that help users work smarter.

With Oracle’s roadmap focused on embedded intelligence and conversational UI, your Redwood extensions can now become part of an AI-powered Fusion experience that’s proactive, human, and scalable. ✔