How to automate email into Infor LN
Short answer
To automate email for Infor LN, read shared mailboxes with Microsoft Graph or IMAP, extract structured data from the message and attachments, match it to LN masters, and post it to LN as a draft through ION or the LN API. A human reviews low-confidence items before anything is committed.
A lot of LN order entry is really email handling. Customer purchase orders arrive as PDFs, suppliers reply with delivery dates in the body of an email, and invoices land in an accounts payable inbox. Someone retypes each into a session, and the same person chases the exceptions.
Automating this is realistic, but the failure mode is silent bad data in the ERP. The reliable approach treats email as untrusted input: classify it, extract fields, validate against LN, and only then create a draft transaction for review or auto-post when confidence and rules allow.
This page explains how to connect mailboxes to LN's integration points, which flows pay off first, and the guardrails that stop a misread email from becoming an order.
What teams use it for
Customer PO to sales order draft
Read POs from a sales inbox, match customer, ship-to and item numbers or customer part numbers, and create a draft sales order for order desk review.
Supplier acknowledgements
Parse replies to purchase orders for confirmed dates and quantities, and propose updates to the order lines for a buyer to accept.
RFQ intake
Detect requests for quotation, extract line items and requested dates, and create a quotation shell for sales to price.
Shipping notices and certificates
Attach advance ship notices and certificates of conformance to the receipt or purchase order they belong to, based on order numbers in the message.
Customer status inquiries
Recognise 'where is my order' emails, look up status read-only, and draft a reply for customer service to send.
Infor LN integration surface
- Infor ION API Gateway: REST access with OAuth 2.0 for creating or updating LN documents where API sets are available, usually the cleanest write path.
- ION Connect and BODs: inbound BODs such as sales order or purchase order documents can be posted into LN through an ION connection point, depending on version.
- LN DAL and business interfaces, which enforce LN validation rules and are preferred over any direct database write.
- LN Exchange schemes for file-based import of sales orders and similar data on older or non-ION installations.
- Infor Document Management or attachment features to link the original email and PDF to the LN transaction for traceability.
How it works
- 1
Inventory the mailboxes and message types
Pick one or two shared inboxes and sample a few hundred real messages. Note formats, languages and the proportion that are structured enough to automate.
- 2
Connect to the mailbox
Use Microsoft Graph for Exchange Online or IMAP for on-prem mail, with a service account scoped to those folders. Move processed messages to a folder instead of deleting them.
- 3
Classify and extract
Classify the message type, then extract header and line fields from body and attachments. Use deterministic parsers for known layouts and a language model for free-form text, always returning confidence per field.
- 4
Match to LN masters
Resolve customer, supplier, item and order references against LN data, including customer-item cross references. Fuzzy matches go to review, not straight through.
- 5
Create a draft in LN
Post through ION or a business interface into a status that does not trigger downstream actions until reviewed, and attach the original email.
- 6
Review and learn
Show reviewers a side-by-side of the source and extracted data. Track corrections to tune rules and decide which flows may be auto-posted.
Example flows
Inbound PO from a repeat customer
Trigger: An email with a PDF purchase order arrives in the orders inbox.
- The service classifies it as a customer PO and extracts header and lines.
- Customer and customer part numbers are matched to LN, and prices are compared to the contract price list.
- If all matches are exact and prices agree, a draft sales order is created; otherwise it is flagged.
- The order desk reviews, corrects any field, and releases the order.
Result: Order entry time drops for routine orders, and mismatches are caught before they become shipments.
Supplier delivery date reply
Trigger: A supplier replies to a purchase order email with new dates.
- The service finds the PO number in the subject or body.
- It extracts line numbers and revised dates, and compares them with the open lines.
- A proposed update is shown to the buyer, who accepts it; LN is updated through the API.
Result: Promise dates in LN stay current without buyers retyping supplier emails.
Keeping data on-prem
Email automation can run fully on-prem. Read from your own mail server or Exchange, run extraction with a locally hosted open-weight model on your GPUs, and write to LN through ION or business interfaces inside the network. This keeps customer POs, pricing and drawings from leaving your environment, which matters for aerospace, defense and electronics programs with controlled data. Cloud mail is fine if the extraction and LN stay inside your perimeter and only the mailbox access crosses it, using a tightly scoped Graph application permission.
Security checklist
- Treat email content as untrusted: strip active content, scan attachments and never execute instructions found in message text.
- Use a dedicated mailbox permission scoped to specific folders, not full tenant access.
- Validate every extracted value against LN masters and business rules before creating a draft.
- Keep the original email and attachment linked to the LN document for audit.
- Guard against spoofing: check sender domains and require review for first-time senders or changed bank details.
Ways to get it done
Generic intelligent document processing product
Effort: Weeks to configure, plus per-document licensing.
Good when document layouts are standard and you only need data extraction handed to a flow.
Build in-house on Graph, an extraction library and ION
Effort: Several weeks per flow, and ongoing tuning.
Fits teams with development capacity and a limited number of document types.
Partner build such as Netray
Effort: A pilot on one inbox, expanding by document type.
When extraction should run on a private model and tie into LN review queues and approvals.
Pitfalls
- !Auto-posting from day one; start with drafts and move to auto only for flows with proven accuracy.
- !Ignoring customer-item cross references, which is why many PDF items fail to match.
- !Not handling revisions, where a second email changes a PO already entered.
- !Losing the audit link between the email and the LN order, which makes disputes harder.
Frequently asked questions
Can Infor LN read emails itself?
Not in a way that understands order content. LN can send mail, but reading and interpreting inbound email into transactions needs an external service that feeds LN via ION, a business interface or an import scheme.
How accurate is automated extraction?
It depends heavily on document variety. Repeat customers with consistent layouts can reach high straight-through rates, while free-form emails need more review. Measure on your own samples and set thresholds per field before enabling automation.
Should we let it create orders without review?
Only for narrow cases proven over time, such as repeat customers with exact matches and unchanged prices. Everything else should create a draft that a person releases, because a bad order is costly to unwind.
Does this work with older Baan or LN 6.1 setups?
Often yes, through file-based Exchange schemes or older import interfaces rather than ION APIs. Real-time behaviour and error handling are less convenient, so plan a more structured review queue.
Where does the AI model run?
It can run on your own servers using an open-weight model, which keeps documents inside your network. A hosted model is possible if your policy allows it, but for controlled data we generally recommend private hosting.
Want Email automation working on your Infor LN?
A free 30 minute assessment: we map the integration path for your version and setup, and where it should run.
Get a free assessment