# How to automate email in Infor VISUAL

> Source: https://netray.co/integrations/infor-visual-email-automation
> Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems.
> Section: Integrations

## Short answer

Email automation for Infor VISUAL reads shared inboxes, classifies each message, extracts the fields, and creates draft customer orders, quotes or PO updates for a person to review. Posting should go through the API Toolkit or a staging area, never straight into tables.

## System

Infor VISUAL

## Capability

email-automation

## Intro

In a typical VISUAL shop, a large share of orders and changes arrive by email. Customers send purchase orders as PDFs, suppliers reply with new dates in the body text, and engineering sends revision notices. Someone retypes each one into VISUAL, and errors appear in part IDs, quantities and dates.

The goal is not to remove the person but to move them from typing to checking. A pipeline reads the mailbox, pulls out the order number, part numbers, quantities, prices and dates, matches them against VISUAL master data, and offers a pre-filled draft with the doubtful fields highlighted.

How far you can go depends on how VISUAL is licensed and hosted. With the API Toolkit you can create objects under business rules. Without it, a staging table and a human keying step is still a large time saving.

## Use Cases

### Customer PO to draft customer order

**Description:** A PDF purchase order is parsed into header and lines, customer part numbers are cross-referenced to VISUAL part IDs, and a draft order is prepared.

### RFQ to quote draft

**Description:** Requests for quote are logged with the requested parts, quantities and dates, and a quote shell is created for estimating.

### Supplier acknowledgement

**Description:** Replies on open POs are matched to PO lines and proposed due date changes are queued for the buyer.

### Shipping and certificate requests

**Description:** Customer requests for a packing list or certificate are routed to the right person with the order already looked up.

### Outbound order confirmations

**Description:** Confirmations and shipment notices are generated from VISUAL data and sent with a review step for exceptions.

## System Surface

- Customer order and quote tables: CUSTOMER_ORDER and CUST_ORDER_LINE style structures hold the targets, but names and required fields vary by version, so confirm them in your schema.
- Cross-reference data: customer part numbers, vendor part numbers and unit of measure conversions in VISUAL are the lookup tables that make extraction reliable.
- VISUAL API Toolkit: lets an integration create or update orders under application rules where your version and license allow it.
- Staging approach: a separate table or queue of proposed transactions that a person approves before posting.
- Mail access: Exchange Online or on-premises Exchange through Graph or IMAP, plus VISUAL's own email-out features for confirmations.

## How It Works

### Pick a dedicated inbox

**Detail:** Route orders, RFQs and supplier replies to shared mailboxes so the system does not read personal mail.

### Classify each message

**Detail:** Decide whether it is an order, RFQ, change, acknowledgement or noise. Keep an 'unknown' bucket for a person.

### Extract fields

**Detail:** Use layout-aware extraction and a language model for the body text. Capture a confidence score per field.

### Match to VISUAL master data

**Detail:** Find the customer, the part (through cross-references) and the unit of measure. Flag a mismatch rather than guessing.

### Present a review screen

**Detail:** Show the email next to the draft, with low-confidence fields marked and price differences from the contract shown.

### Post and learn

**Detail:** On approval, post through a supported interface. Store corrections so recurring customers need less review.

## Example Flows

### Repeat customer PO

**Trigger:** A PDF purchase order arrives in the orders mailbox.

**Steps:**

- The pipeline identifies the sender as a known customer.
- Lines are extracted and customer part numbers are mapped to VISUAL parts.
- Prices are compared with the last order and a difference is flagged.
- A clerk reviews the draft and approves it.

**Result:** The order is entered in minutes, with the only attention spent on flagged lines.

### Supplier date change by email

**Trigger:** A vendor writes that two PO lines will ship later.

**Steps:**

- The sender and PO number are matched to open lines.
- New dates are extracted and compared with current due dates.
- The buyer sees a proposal listing each line.
- Approved dates are updated and the planner is notified of affected jobs.

**Result:** Due dates reflect reality without anyone retyping the message.

## On Prem

Email content often includes drawings, pricing and customer part data, so many VISUAL shops want extraction to happen inside their own environment. Run the classifier and extraction models on a local server with open-weight models, keep mailbox access through your own Exchange or a Graph app limited to specific mailboxes, and store drafts in a database next to VISUAL. Nothing needs to leave your network except the mail itself. If an outside service is ever used for OCR on scans, be explicit about what documents it sees, and exclude controlled customers and drawings from it.

## Security

- Limit mailbox access to named shared inboxes with a scoped application permission.
- Scan attachments for malware before parsing and treat every email as untrusted input.
- Defend against prompt injection in email bodies by never letting extracted text trigger actions.
- Require human approval for new customers, new bank details and price changes.
- Keep original emails linked to the VISUAL record for audit.

## Options

### Generic document-processing tool

**Fit:** Works for standard layouts and low volumes, with a manual step to enter the data in VISUAL.

**Effort:** Days to weeks

### Build in-house

**Fit:** Possible if you have developers who know the VISUAL order structures and can maintain extraction rules.

**Effort:** Many weeks

### Partner build with Netray

**Fit:** Appropriate for on-prem extraction models, customer part cross-referencing and a review screen with controlled write-back.

**Effort:** A few weeks per document family

## Pitfalls

- Treating one extraction accuracy figure as applying to every customer layout.
- Auto-posting orders with unit of measure or price mismatches.
- Forgetting duplicate emails and forwarded chains, creating double orders.
- Reading personal mailboxes instead of shared ones.
- Writing straight to tables and skipping VISUAL business rules.

## Faq

### Can Infor VISUAL read orders from email on its own?

**Answer:** Not in a general way that I know of. VISUAL can send email, but turning inbound messages and PDFs into orders needs an extra layer for reading, extraction and matching. Some partners sell add-ons, and the options vary by version, so check before you build.

### Should the system post orders automatically?

**Answer:** Start with review. After a customer's layout has been right for many orders, you can auto-post only the lines where every field is high-confidence and prices match. Keep new customers, price changes and unknown parts for a person.

### How are customer part numbers matched?

**Answer:** Through the cross-reference data in VISUAL, supported by a learned mapping from past corrections. When a part cannot be matched with confidence, the draft shows the closest candidates and asks a person to pick, rather than inventing a part ID.

### Does this need the cloud?

**Answer:** No. Extraction can run on a local GPU server using open-weight models, with the mailbox reached through your own Exchange. A cloud service may be simpler for small volumes, but regulated customers often prefer to keep drawings and pricing in-house.

### How do we avoid duplicate orders?

**Answer:** Store a fingerprint of each message and the customer PO number, and check both before drafting. If the same PO number already exists in VISUAL for that customer, the pipeline treats the email as a change or duplicate and asks for review.

## Search terms

infor visual email automation, automate purchase orders email infor visual, visual manufacturing order entry automation, email to customer order erp, infor visual rfq automation, extract po from pdf email erp, infor visual edi alternative email, automate order entry job shop erp, infor visual api toolkit orders

---
Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json