Every supplier sends a different Excel file. Different columns, different units, different color names. At wholesalers and importers with a hundred suppliers or more, two or three people often work full time on processing those files by hand. And it does not stop at data entry: translations, descriptions and quality checks usually follow separately and manually, from scratch again with every new file.
SDM Portal solves that with a visual flow editor. You drag the columns of a supplier file alongside the fields of your PIM, whether that is Ergonode, Akeneo or Pimcore, and connect them with transform blocks: value mapping, splitting, merging, unit conversion. For the awkward rules you describe in plain language what you want, and the system compiles that once into a fixed rule with a confidence score. The same flow holds a translation block that recognizes the source language and translates into the right locale, and a block that generates a description or a short description out of several source columns. Processing and enrichment happen in one step.
The AI does its thinking while the mapping is being built, not on every run. After that it runs on pure TypeScript: no model calls, every run identical and verifiable. The system also learns per customer environment: every profile you approve is remembered, so the next supplier file starts with better suggestions. AI here speeds up setting up and improving a mapping, it is not the basis of the mapping itself.
Every retailer or wholesaler with a lot of suppliers has the same problem: the same Excel clean-up job every month, and then separate work on top of it to get translations and descriptions in order. With SDM Portal you build the mapping once, enrichment included. After that it runs itself, on every new file from that supplier, automatically or on a schedule, and it gets a little better with every profile you approve. If a supplier changes the layout of its file in the meantime, the system spots the difference, puts the profile back into the workflow and warns the owner, so wrong data is never pushed through unnoticed.
Every change is in an audit log, a production run or the deletion of a profile can be tied to mandatory 2FA, and customers in the same installation are separated from each other at database level. For a team that shares supplier data across several people, that is the difference between a loose Excel culture and a process you can rely on, even while that process improves itself along the way.