# How to automate email in Infor M3 > Source: https://netray.co/integrations/infor-m3-email-automation > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Integrations ## Short answer Automate email in Infor M3 by reading shared inboxes, classifying and extracting each message, and creating draft transactions through M3 API programs or ION, with a review queue before anything posts. Start with customer orders and supplier confirmations, where the data is structured enough to validate against M3 masters. Keep write steps behind human approval until error rates are known. ## System Infor M3 ## Capability email-automation ## Intro A surprising amount of M3 order entry still starts as an email: a customer PO attached as a PDF, an Excel order sheet from a distributor, a supplier replying with new dates, a forwarded RFQ. Someone reads it, finds the customer and items in M3, and types it into the order entry program. That is slow and error-prone, and it concentrates knowledge in a few people. Email automation does not mean a robot that emails customers. It means reading what arrives, matching it to M3 masters (customers, items, suppliers, open orders), and producing a proposed transaction that a person can accept. The value comes from matching quality and review design, not from the language model alone. M3 offers APIs and ION for the integration, and Infor OS has document and workflow tooling, but none of it parses unstructured mail for you. You add that layer, and decide how much of it can run on-prem. ## Use Cases ### Customer PO to customer order **Description:** Extract customer, ship-to, item numbers, quantities and requested dates from an emailed PO, match to M3 masters and create a draft customer order for review. ### Supplier order confirmation updates **Description:** Read supplier replies, find the purchase order and line, extract confirmed dates and quantities, and propose an update to the PO line. ### RFQ intake and routing **Description:** Recognise a request for quote, create an inquiry record or a task, and route it to the right sales owner with extracted items and quantities. ### Invoice mail to accounts payable **Description:** Route supplier invoices from a shared AP mailbox into the invoice capture process, flagging duplicates by supplier and invoice number. ### Shipping and delivery notices **Description:** Parse carrier notices and advance shipment or delivery status, flagging mismatches against the planned delivery. ### Customer service triage **Description:** Classify inbound requests (status, change, return, complaint), attach order context from M3 and assign to the right queue. ## System Surface M3 API (MI programs) for creating and updating customer orders, purchase orders and business partner records. Program names such as OIS100MI-style for orders vary by version, so confirm in the API metadata. ION API gateway and ION Connect for secure access and for emitting or consuming BODs, with an OAuth 2.0 service account created in Infor OS. M3 master data tables and APIs (customers, items, supplier-item, customer-item cross references) used for matching: customer-item and supplier-item cross references are the key to resolving partner part numbers. Infor Document Management (IDM) or the attachment features, to store the original email and attachments linked to the order or PO. M3 Enterprise Collaborator (MEC) and EDI partner agreements for customers who already send structured EDI; use these for clean volumes and mail automation for the long tail. ## How It Works ### Inventory the mail flows **Detail:** Sample a few months of the sales, purchasing and AP mailboxes. Count by type, sender and format. Pick the one or two flows with the best volume and the most regular layouts. ### Connect the mailbox safely **Detail:** Use a dedicated shared mailbox with Microsoft Graph or IMAP access, not an individual's inbox. Process messages idempotently using message IDs and mark them with a state. ### Classify and extract **Detail:** Classify the message type, parse body and attachments (PDF text or OCR, Excel), and extract fields into a typed schema. Prefer rules for known customers and a model for the long tail. ### Match to M3 masters **Detail:** Resolve the sender to a customer or supplier, map partner part numbers via customer-item or supplier-item cross references, and verify units and currencies. Anything unmatched goes to review. ### Create drafts and review **Detail:** Show the email, extracted data and proposed transaction side by side. A person confirms, corrects and releases. Only then does the service call the M3 API. ### Measure and widen **Detail:** Track straight-through rate, correction rate by field and customer, and time saved. Expand automation customer by customer as accuracy proves out. ## Example Flows ### Emailed PO becomes a draft customer order **Trigger:** A PDF purchase order arrives in the orders mailbox. **Steps:** - The service extracts text, identifies the customer from sender domain and PO header. - It maps customer part numbers to M3 items and checks quantities and requested dates. - It prepares a proposed order with warnings for price differences and unknown items. - A sales administrator reviews, fixes one line, and approves; the service creates the order through the API and stores the email as an attachment. **Result:** Order entry time drops sharply on the accepted flows, and every order links back to its source email. ### Supplier date change **Trigger:** A supplier replies to a PO with revised delivery dates. **Steps:** - The service matches the PO number in the subject or body to an open PO. - It extracts line references and new dates. - A buyer reviews a diff of old and new dates and accepts or rejects it. - On accept, the line date is updated through the PO API and a note records the source. **Result:** Fewer stale dates in MRP, and buyers deal only with real changes. ### RFQ triage **Trigger:** An RFQ lands in the shared sales mailbox. **Steps:** - The message is classified as an RFQ and customer details are extracted. - Requested items are matched to M3 items where possible. - A task is assigned to the account owner with the extracted lines and original files. **Result:** Quotes start the same day with the right owner, and no request sits unread. ## On Prem Email content often includes pricing, drawings and customer data, so on-prem processing is a reasonable default for defense, aerospace and electronics customers. Run the parsing and extraction on a local open-weight model on your own GPU server, keep M3 API calls on the internal network, and store attachments in IDM or a local file store. If you use Microsoft 365 mail, connect through Graph from inside your network and keep processed content local. The extraction model, matching rules and review UI can all run without any external API. ## Security - Treat email as untrusted input: strip active content, scan attachments, and never let message text change system prompts or API parameters directly. - Check sender identity (domain, SPF and DKIM results) before auto-matching to a customer; flag lookalike domains. - Limit the M3 technical user to the specific order and PO programs, and require review before writes. - Store the original message and attachments with the M3 record for audit, and retain per policy. - Log every extraction, match and human correction; keep personal data minimal in logs. ## Options ### Existing intelligent document processing product with M3 connector **Fit:** High volumes of regular documents where vendor support is valued. **Effort:** Weeks to a few months, with per-document pricing. ### Build in-house using Graph, a model and M3 APIs **Fit:** Teams with development capacity and a limited set of flows. **Effort:** Two to four months to a dependable release. ### Partner build such as Netray **Fit:** On-prem or air-gapped requirements, mixed formats, and a need for review UI and measurement. **Effort:** A few weeks for a scoped pilot on one flow. ## Pitfalls - Auto-posting orders without a review stage, which turns a mis-extracted quantity into a shipped error. - Ignoring customer-item cross references, so partner part numbers fail to match and everything lands in review. - Processing a personal mailbox, which mixes unrelated mail and complicates access control. - Treating all PDFs as text; scanned documents need OCR and a confidence threshold. - Missing duplicate detection, so a resent email creates a second order. ## Faq ### Can Infor M3 read emails natively? **Answer:** Not as an order entry feature that I would rely on. M3 provides APIs, EDI tooling and ION, but parsing free-form emails and attachments is done by a layer you add. That layer then creates transactions through M3 APIs or ION. ### How do we avoid wrong orders? **Answer:** Use a review step at first, validate against masters (customer, item, unit, currency) and show warnings for price or date differences. Move to straight-through only for specific customers whose accuracy is proven over weeks of data. ### What about EDI customers? **Answer:** Keep EDI for customers that support it, since structured messages are more reliable. Use email automation for the long tail that sends PDFs and spreadsheets. Both can create the same order through the same API path. ### Can this run fully on-premises? **Answer:** Yes. Mail access, OCR, extraction with a local open-weight model, matching and the review UI can all run on your servers. Only mail transport itself is external. Accuracy depends on your document mix, so test on your own samples. ### Where do we store the original email? **Answer:** Attach it to the order or PO using IDM or the attachment feature in your version, or in a document store with a link on the M3 record. The goal is that an auditor can go from an order line to the source message in a click. ## Search terms infor m3 email automation, automate purchase orders email infor m3, infor m3 order entry automation, infor m3 inbound email to customer order, infor m3 rfq email automation, email to erp automation infor m3, infor m3 supplier confirmation automation, infor ion email integration, infor m3 shared inbox automation, ai email processing infor m3 --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json