AI Agent

+

JD Edwards

Turn Incoming Order Emails Into JD Edwards Sales Orders — Automatically

AtomIQ AI Agent reads an inbox, understands the order details, and creates a Sales Order directly in JD Edwards through E1 Orchestrator. No manual entry, no custom CNC scripting, works with the LLM provider you already trust.
See It Built Live
See How It Works

How the Workflow Actually Runs

This is the real workflow architecture, not a simplified diagram.

✉️

Gmail

Trigger

→

⑂

Switch

Route by content

→

🤖

AI Agent

Analyzes email

→

{ }

Inline Transform

Maps to ERP fields

→

◆

E1 Orchestrator

Creates Sales Order

1

Email arrives

A customer sends a purchase or sales order request to a monitored inbox.
2

AI Agent analyzes the content

The connected LLM reads the email and extracts order details — items, quantities, customer, requested dates.
3

Inline transform maps the data

Extracted fields are mapped into the structure JD Edwards expects, correcting formats and resolving references.
4

E1 Orchestrator creates the Sales Order

AtomIQ calls JD Edwards’ native Orchestrator REST interface to create the order directly — no screen-scraping, no CNC customization required.

Manual Entry vs. AI Agent

TaskManual ProcessAtomIQ OCR Pipeline
Reading order emails Staff reads and interprets each email individually AI Agent reads and extracts order details automatically
Entering the Sales Order Manual keying into JD Edwards Created automatically via E1 Orchestrator
Error rate Higher — manual re-typing introduces mistakes Lower — data mapped directly from source, no re-typing
Time per order Minutes per order, queued behind other tasks Near-instant, runs as soon as the email arrives

Works With the LLM Provider You Trust

Not locked into a single AI vendor. Configure the provider that fits your security, cost, or data residency requirements.

Anthropic

OpenAI

Azure OpenAI

Google Gemini

AWS Bedrock

DeepSeek

Mistral

Grok

Ollama (self-hosted, private)

Other Real AI Agent + JD Edwards Workflows

This isn’t a one-off demo — the same platform runs other document and order automation workflows.

Purchase Order OCR Processing

Azure Document Intelligence extracts structured data from PO/invoice documents pulled from a file location, then routes it into JD Edwards via an automated pipeline.

Bill of Lading E-Signature
JD Edwards E1 REST data populates a shipping document, routed through DocuSign for electronic signature capture — built as a working demo template for warehouse and shipping workflows.
Multi-Step Order Validation
Inline JS transforms and REST calls validate and enrich order data before it’s committed to JD Edwards, reducing bad data reaching the ERP.
The Bill of Lading example above is a demonstration template built to show the DocuSign + JD Edwards integration pattern — not a completed client deliverable. Sample data only.

Frequently Asked Questions

Yes. AtomIQ AI Agent monitors an inbox, uses an LLM to analyze incoming order emails, extracts the order details, and creates a Sales Order in JD Edwards through the E1 Orchestrator interface, without manual data entry.
AtomIQ supports Anthropic, OpenAI, Azure OpenAI, Google Gemini, AWS Bedrock, DeepSeek, Mistral, Grok, and self-hosted models via Ollama, so you can choose the provider that fits your security and cost requirements.
No custom CNC scripting is required for the integration itself. AtomIQ connects to JD Edwards through the E1 Orchestrator REST interface and a visual workflow designer, with an inline transformation step to map extracted data into the correct JDE fields.
Yes. AtomIQ also supports OCR-based document processing using Azure Document Intelligence, extracting structured data from purchase order or invoice documents pulled from a file location or blob storage.

See This Running on Your Own Inbox

Book a session and we’ll build a working version against a real workflow, not a slide deck.
Book a Demo