Use case

Automated Checks for Product Documents and Certificates

Automated pre-checks catch missing fields, wrong dates, duplicates and mismatched PDF files before documents and certificates reach review and the product passport.

automated document checksproduct data quality checkscertificate validation software

Problems this page solves

Documents and certificates arrive with missing fields and wrong dates.
The same certificate gets added several times with a different number or issuer.
A PDF file does not always match the product or record it was added to.

What the team gets

Warnings about gaps, dates and duplicates as soon as a file is added or imported.
The file name and readable PDF text are compared with the record data.
A person reviewing the document always makes the approval decision.

Next step

See this workflow on a realistic product demo

We will show the product flow, public page, automated checks and API angle that best matches this use case.

Deep dive

A more detailed view of the workflow behind this search intent

What the automated pre-checks look at

Every time you add a document or certificate, import data from a CSV file or send it through the API, DPPC runs a set of rules that check data quality. The result is a list of specific warnings and suggestions, such as a missing certificate number, an expiry date earlier than the issue date or a file whose name does not match the product. Errors show up right away instead of during review or after the product passport is published.

Certificates: data, dates and duplicates

For a certificate, the system checks whether the issuer and number are filled in, whether the expiry date comes after the issue date, whether the certificate is already expired when saved or expires soon, and whether its status matches its dates. It also detects a certificate with the same name saved earlier under a different number or issuer. It warns when the PDF file name does not match the certificate name or number, or when the file is suspiciously small and may be incomplete.

Product documents: does the file match the record?

For product documents, DPPC flags a document with the same title and type added a second time, very short titles and files whose names contain neither the SKU nor the product name. If the PDF has a readable text layer, the system also checks whether the document mentions the product and its own title, and whether it contains a different SKU, GTIN, batch number or serial number than the product record. It is a simple way to catch a document attached to the wrong product.

Data imports: products, materials and suppliers

The same rules apply to CSV imports. For products, the system flags a missing brand, model or country of origin, a description that is too generic, an unusual GTIN format and empty industry fields. For materials, it detects shares that add up to more than 100 percent, suspiciously low values, mass without a unit, a missing supplier and a country of origin that differs from the supplier country. For suppliers, it points out missing contact details or country and a VAT ID already assigned to another supplier.

Warnings, suggestions and a human decision

The checks never reject anything on their own. A document or certificate with warnings goes to review with the status "needs review", and the person in charge sees what each warning is about. If the readable PDF text contains a certificate number or issuer and that field in the record is empty, the system can suggest it. Every result also includes a confidence score. Through the API, the same checks are available in a preview mode that returns warnings without saving any data.

What the automated checks do not do

We are upfront about this, because trust in the process depends on it. The checks are predictable, transparent rules, not an artificial intelligence model. They do not recognize text in scans, and they only read content from PDF files that have a plain text layer. They also do not judge whether a document is correct in substance. Their job is to catch gaps, inconsistencies and mistakes before a person looks at the document, and the approval decision is always made by the person reviewing it.

FAQ

Frequently asked questions

Do the automated document checks in DPPC use artificial intelligence?

No. They are a set of transparent rules that check data, dates, duplicates, file names and readable PDF text. Every warning has a clear cause, and the same document always produces the same result.

Do the checks work with scanned documents?

For scans, the system checks the data entered in the record, the file name, the file size and duplicates, but it does not read content from the page image because there is no text recognition (OCR).

Does a warning block a document from being saved?

No. The document is saved and marked as needing review. Publishing the passport is only blocked by the readiness score, for example when a certificate is rejected or expired.

Can I check data through the API before importing it?

Yes. The API has a preview mode for products, suppliers, certificates and documents that returns warnings and suggestions without saving any data.

What is the confidence score on a check result?

It is a simple indicator of how many signals from the document match the record and how many warnings were found. It helps decide what to review first, but it does not replace the decision of the person in charge.