Upload the file you got from your supplier in CSV, XML or Excel. Pobo pulls the products out of it, looks up what the file is missing and fills in parameters and descriptions.
It does the copying, the searching and the checking. You only approve the result.
Upload what the supplier sent you. Pobo recognises the format by its content and finds the products in it.
Pobo suggests which column holds the name, the price or the EAN. You only confirm it.
Brand, category and parameters the file lacks get searched across several AI models and compared.
The next file from the same supplier maps itself. Feeds are stable per supplier, so guessing again would only add noise.
A missing protocol gets fixed, duplicates in one row are merged and a value that is not an address at all is thrown away instead of saved as a broken photo.
A file with Parameter name 1 and Parameter value 1 is read as pairs, not as two useless columns.
Copying is slow and a plain import brings only what the file has. AI listing adds the missing facts and checks them.
The batch validates first and stops. Until you approve it, nothing is created and nothing is charged.
A row with no name or a duplicate code is written down with its state, never dropped. The count in the batch always matches the file.
What several models agree on is treated as reliable. What only one of them claims stays marked, so an invented parameter does not end up in your catalogue.
Before the batch starts you see how many rows will go through, which ones have a problem and what it will cost.
From a supplier file to finished products in four steps.
CSV, XML or Excel. We recognise the format by its content, so a feed saved as .txt works too.
Pobo suggests which column holds the name, the price or the EAN. You only check it and fix what you want.
Brand, category and parameters get searched across several AI models and compared against each other.
In the same format you uploaded, or straight into your product pages in Pobo.
CSV, XML or Excel. We recognise the format by its content, so a feed with a wrong or missing extension works as well. The file does not have to have a prescribed structure, that is what the mapping step is for.
They become product parameters named after their own heading. Nothing from the file is lost, which matters for supplier feeds where most of the value sits in columns like Colour, Material or Waterproofing.
It still goes through. Larger wholesalers synchronise prices from their own warehouse, so a price list without prices is a normal state. A product we do not know the price of is simply never made visible on its own.
Several models search independently and the results are compared. Agreement across sources counts as confidence, a single source stays marked as unverified. We would rather leave a field empty than fill it with something invented.
Only if you want. Products are created in Pobo and you decide what happens next: take the data as a file, or publish the descriptions to your e-shop the same way you do today.
In thirty minutes Lenka runs a real file of yours through AI listing and shows you what Pobo fills in. If you would rather try it on your own, you can start right away.
Book a call