Tuesday, 13 January 2026

How We Built an AI-Powered Digital Workforce for JD Edwards

 

How We Built an AI-Powered Digital Workforce for JD Edwards

Transform your JDE operations with AI agents that audit, analyse, and automate


The Challenge: Buried in Data, Starved for Insights

If you manage JD Edwards EnterpriseOne, you know the paradox: your ERP contains millions of transactions and decades of business logic, yet extracting meaningful insights feels like archaeology. Running integrity reports, auditing module configurations, researching Oracle documentation, and generating compliance reports consumes countless hours of expert consultant time.

What if your JDE system could audit itself? What if an AI agent could run integrity reports, interpret the results, identify issues, recommend fixes, and generate professional audit reports - all in minutes instead of days?

That's exactly what we've built at Fusion5.


Introducing the JDE Digital Workforce

The JDE Digital Workforce is an AI-native platform that deploys specialized agents to work alongside your finance and operations teams. These agents understand JD Edwards at a deep technical level - they know your table structures, integrity reports, business functions, orchestrations, and can execute complex multi-step audits autonomously.

What Can It Do?

Our Digital Workforce agents can:

  • Audit any JDE module - AP, AR, GL, Sales, Inventory, Manufacturing, Procurement
  • Execute and interpret integrity reports - Run R04701, R41543, R09701, retrieve PDFs, analyse results
  • Identify data quality issues - Orphan records, negative inventory, duplicate vendors, unposted entries
  • Research Oracle enhancements - Find relevant features from recent releases that address your issues
  • Generate professional deliverables - Branded audit reports with charts, findings, and remediation steps
  • Discover orchestrations - Search your orchestration library for automation opportunities

All of this happens conversationally. You simply ask: "Run a comprehensive Accounts Payable audit" and the Digital Workforce handles the rest.


The Building Blocks: How It Works

1. Model Context Protocol (MCP)

The foundation of our solution is Anthropic's Model Context Protocol (MCP) - an open standard that allows AI models to securely connect to enterprise systems.

Think of MCP as a universal adapter. Instead of building custom integrations for each AI model and each data source, MCP provides a standardized way for:

  • AI models (Claude, GPT, etc.) to discover available tools and data
  • Servers (your JDE environment) to expose capabilities through a standard interface
  • Clients (our agents) to invoke operations without knowing implementation details

2. JDE AIS MCP Server

Our JDE AIS MCP Server wraps the JD Edwards AIS (Application Interface Services) REST API and exposes it to AI agents via MCP. It provides 43+ tools:

Tool CategoryExamplesPurpose
Data Serviceais_data_service_v2Query any JDE table (F0411, F03B11, F4111)
Form Serviceais_form_service_executeExecute JDE applications (P0411, P4210)
Report Executeais_report_executeais_report_statusRun UBEs and check status
BSFN Executeais_bsfn_discoverais_bsfn_executeCall business functions
Knowledgeais_table_knowledgeais_column_knowledgeExplain JDE tables and fields

🔴 Current Status: AIS Server is currently offline (f5play.fusion5.cloud unreachable)

3. JDE Orchestrations MCP Server

The Orchestrations MCP Server provides access to JDE Orchestrator capabilities:

  • Search orchestrations - Find orchestrations by keyword
  • Execute orchestrations - Run orchestration workflows
  • Get UBE output - Retrieve PDF output from batch jobs (via orch_getUBEOutput)
  • Discover integrations - Explore available orchestrations in your environment

This enables agents to not only analyse data but also discover and recommend automation opportunities.

4. Expert Agents

We've created 11 module-specific expert agents, each with deep knowledge of a JDE functional area:

  • GL Expert - General Ledger audits
  • AP Expert - Accounts Payable analysis
  • AR Expert - Accounts Receivable reviews
  • Inventory Expert - Inventory integrity checks
  • Sales Expert - Sales order analysis
  • Procurement Expert - Purchase order audits
  • Manufacturing Expert - Work order reviews
  • Fixed Assets Expert - Asset tracking audits
  • Payroll Expert - Payroll validation
  • Address Book Expert - Master data quality
  • Real Estate Expert - Property management audits

Each expert agent knows:

  • Primary tables to query (e.g., F0411 for AP, F4211 for Sales)
  • Integrity reports to execute (e.g., R04701 for AP, R41543 for Inventory)
  • Status codes and their meanings
  • Common data quality issues to check for
  • AAI (Automatic Accounting Instructions) to validate
  • Best practices for the module

5. The Audit Controller

The Audit Controller is the orchestrator that ensures consistency across all audits. It defines 22 mandatory sections that every audit must include:

  1. Preface - About JDE, about AI, about the audit
  2. Technical Landscape - Environment configuration
  3. Environment Fundamentals - F0010, F0006, F0092 queries
  4. Integrity UBE Execution - Run and interpret reports
  5. Executive Summary - Key findings at a glance
  6. Module-Specific Analysis - Deep dive into the module
  7. User & Data Currency Analysis - Who's using the system, how fresh is the data
  8. Status/Code Analysis - Breakdown by status codes
  9. Aging/Period Analysis - Historical trends
  10. Entity Analysis - Company, business unit, supplier/customer breakdown
  11. Data Quality Checks - Duplicates, orphans, anomalies
  12. Audit Findings Summary - Categorized by severity
  13. Recommendations - Prioritized action items
  14. Orchestration Discovery - Available automation
  15. Data Visualisation - Charts using Fusion5 brand colors
  16. Findings with Remediation - How to fix each issue
  17. Release Enhancement Opportunities - Oracle features that can help
  18. Fusion5 Services - How we can assist
  19. Document Generation - Final deliverable creation

This ensures that whether you're auditing Accounts Payable or Inventory, you get a consistent, comprehensive report.

6. The Audit Completion Agent

Before any audit is finalized, the Audit Completion Agent validates that all 22 mandatory sections are present and complete. If anything is missing, it identifies the gaps and triggers completion.

This acts as a quality gate - no audit leaves our platform incomplete.

7. Skills for Specialized Operations

We've created reusable skills that agents can invoke for specific tasks:

SkillPurpose
jde-aisDirect AIS operations reference
jde-orchestrationsOrchestration execution patterns
jde-batch-job-interpretationRun reports, retrieve PDFs, interpret results
jde-enhancementsResearch release enhancements
fusion5-documentGenerate branded Word documents

Skills provide structured knowledge that agents can reference when performing complex operations.


Real-World Example: Inventory Audit in Minutes

Here's what happens when you ask the Digital Workforce to audit Inventory:

Step 1: Environment Connection

The agent calls ais_get_config to understand your environment:

  • JDE version (e.g., E9.2.9.4)
  • Environment name (JPLAY920)
  • Server details (F5PLAY2)

Step 2: Data Analysis

It queries key inventory tables:

  • F4101 - Item Master (2,578 items found)
  • F41021 - Item Location (5,528 location records)
  • F4102 - Item Branch Plant (3,633 branch records)
  • F4111 - Item Ledger (15,346 transactions)

Step 3: Pattern Recognition

The agent identifies critical issues:

  • ⚠️ 90 items with negative inventory totaling -500M units
  • ⚠️ Soft committed quantity of -499M (stale commitments)

Step 4: Integrity Report Execution

It executes the mandatory integrity reports:

  • R41543 - Item Ledger Integrity (Job 1620 - Complete)
  • R41540 - Item Location Integrity (Job 1621 - Complete)
  • R41550 - Item Ledger to Location (Job 1622 - Complete)
  • R41500 - Item Balance to GL (Job 1623 - Complete)

Step 5: PDF Retrieval and Interpretation

For each report, it:

  1. Retrieves the PDF via orchestration
  2. Decodes and extracts the text
  3. Interprets the results
  4. Explains business impact


Step 6: Comprehensive Report Generation

Finally, it generates a professional audit report with:

  • Executive summary with risk ratings
  • Detailed findings with remediation steps
  • Data visualizations
  • Oracle enhancement recommendations
  • Next steps

Total time: 3-5 minutes (vs. 4-8 hours manually)


Why This Matters: The Business Impact

For Finance Teams

  • Reduce audit preparation time from weeks to hours
  • Improve data quality with continuous monitoring
  • Accelerate period close by catching issues early
  • Strengthen internal controls with documented evidence

For IT Teams

  • Proactive issue detection before users report problems
  • Faster root cause analysis with AI-powered investigation
  • Better utilization of JDE investment by understanding what you have
  • Knowledge preservation as expertise is codified in agents

For Leadership

  • Data-driven decisions backed by real-time insights
  • Risk mitigation through continuous compliance monitoring
  • Cost reduction by automating routine analysis
  • Competitive advantage through AI augmentation

The Technical Advantage: Why MCP Changes Everything

Traditional ERP integration approaches have major limitations:

Traditional ApproachMCP-Powered Approach
Custom APIs per modelOne standard protocol
Hard-coded integrationsDynamic tool discovery
Manual prompt engineeringAgents learn from schemas
Static workflowsAutonomous decision-making
Brittle when systems changeSelf-adapting to capabilities

MCP transforms AI from a chatbot to a colleague. The AI doesn't just answer questions - it actively works with your systems, discovers available tools, and accomplishes complex multi-step tasks.


What's Next: Expanding the Digital Workforce

We're actively expanding capabilities:

Near-Term Roadmap

  • Continuous monitoring agents - Scheduled audits that alert on anomalies
  • Remediation agents - Agents that not only identify issues but fix them
  • Enhanced visualizations - Interactive dashboards from audit data
  • Multi-environment comparison - DEV vs. TEST vs. PROD analysis
  • User training agents - AI tutors for JDE users

Research Areas

  • Predictive analytics - Using historical patterns to forecast issues
  • Natural language queries - "Show me vendors we haven't paid in 90 days"
  • Automated testing - Agents that validate JDE changes before promotion
  • Integration discovery - Mapping data flows across your JDE ecosystem

Getting Started

The JDE Digital Workforce is available now for Fusion5 customers. We can:

  1. Deploy to your environment - Connect to your JDE AIS endpoint
  2. Customize expert agents - Tailor to your specific business rules
  3. Train your teams - Show your staff how to work with AI agents
  4. Deliver ongoing value - Continuous improvement as agents learn

Pricing Models

  • Project-based - One-time audit deliverables
  • Subscription - Ongoing access to the Digital Workforce platform
  • Hybrid - Core platform + custom agent development

The Future is AI-Native

Five years from now, every JDE customer will have a Digital Workforce. The question is: will you be leading the transformation, or catching up?

ApproachDescription
Traditional consultingSend a team for 3 weeks, deliver a report, leave.
Digital WorkforceDeploy agents that work 24/7, learn your environment, and provide continuous insights.

The technology is here. The agents are ready. The only question is: what will you have them do first?


Contact Information

Contact: Shannon Moir | shannon.moir@fusion5.com.au

Try it: Request a demo audit of your JDE environment


Published by Fusion5

Trusted JD Edwards partner across Australia and New Zealand

www.fusion5.com.au | jde@fusion5.com.au