MSF plugs into the systems you already own
Keep your ERP. MSF connects to it. Orders, items, BOMs and stock flow in automatically; confirmations, consumption and quality data flow back. Plus: digitalize any paper-based process flow across your company.
Request a DemoStart a PoCERP integration for manufacturing connects the ERP to the shop floor: production orders, items, BOMs, routings and stock flow down automatically, while confirmations, material consumption, scrap and quality results flow back without double data entry. Meta Smart Factory offers standard connectors for SAP, Microsoft Dynamics, Oracle NetSuite and Odoo, REST APIs or file exchange for other ERPs, and a read-only first phase before anything is written back.
SAP, Dynamics, NetSuite, Odoo and many regional ERPs via standard connectors; anything else via REST API or file exchange.
The integration layer maps custom fields and logic. Customization is normal and does not block the project.
MSF sits alongside the ERP: you keep finance and commerce where they are and add a world-class manufacturing layer.
Inbound, typically production orders, item master, BOMs, routings and stock levels. Outbound, order and operation confirmations, produced and scrapped quantities, material consumption and quality results. The argument is never about that list, it is about which system owns each field and what happens when both are edited. Agree ownership field by field before anyone writes a mapping, because that decision is expensive to reverse.
Yes, if confirmations are to post back. Anything that books a goods movement or closes an operation is a write, and your ERP team will want it done through released interfaces such as BAPIs, OData services or a staging table rather than direct database access. A read-only first phase is reasonable: pull orders, show them on the floor, keep booking by hand until the numbers agree.
Weeks when a named owner on the ERP side, a test system that resembles production, agreed interfaces and broadly correct master data are already in place; longer when they are not. The connector work is rarely the long pole; waiting for change windows, transport approvals and someone to answer which of four item numbers is the real one is. If you cannot name the ERP contact today, plan in months rather than weeks.
Through scheduled CSV or XML drops over SFTP, an ERP-side staging table, or a database view that your ERP vendor or your own ERP team builds and supports; reading application tables yourself is usually refused and can break on an upgrade. What you lose is immediacy, and the lag is whatever interval you schedule, commonly per shift or nightly. The return leg needs an ERP-side import job with an owner and a rule for records it rejects.
They will disagree, and no integration decides who is right. Differences come from bookings made at different moments, scrap recorded on the floor but not yet posted, and movements done outside both systems. What you get is an event trail, so a difference can be traced back to a shift and an order instead of appearing as a lump at stocktake. Name a reconciliation owner and a cadence before go-live.
If the ERP is unreachable but the local server and plant network are running, orders already released stay on the panels and operators keep confirming. Those confirmations are written to the local database, so they survive a restart and are sent once the ERP answers. The practical limit is orders that were never transferred, which cannot be started. Ask whether replayed postings can land in a period your finance team has closed.
No. The process flow designer digitalizes forms, approvals and the routing of work between people, so paper deviation and engineering change processes move onto it well, but it does not change how your ERP prices, plans or posts, and any step that has to update an ERP record still needs an interface behind it. Logic that is really calculation, such as costing rules, belongs where that data lives.
Get a free demo and a Proof of Concept plan tailored to your production.