/ document format / source quality / supplier evidence
When the Supplier Changes the Document Format
Why a shift from formal PDFs to images, edited files, or summaries should trigger source-quality review.
A supplier may send the same information in a different format: PDF becomes screenshot, original certificate becomes cropped image, invoice becomes spreadsheet, bank letter becomes typed email, audit report becomes summary slide. The data may look familiar, but the source quality changed. Reviewers should treat document format changes as evidence events when they affect critical fields.
The first question is what the new format removes. Screenshots remove metadata and page context. Spreadsheets can change formulas or fields. Summaries remove annexes and footnotes. Edited PDFs may hide revision history. A typed email can explain a relationship but cannot replace a formal authorization when payment depends on it. AI may read all formats well, but reading is not verification.
The workflow should compare format and field together. A product photo in chat may be fine for early discussion. A certificate scope page in chat may be too weak for product approval. A bank account typed in an email should not replace a confirmed invoice or bank letter. Source quality follows the decision, not the file extension alone.
Supplier requests can stay narrow. Please send the original PDF with all pages. Please provide the invoice as issued rather than copied fields. Please include the annex page that shows model coverage. These requests tell the supplier exactly what format gap blocks review. They also reduce the chance of receiving another polished but weak file.
The final note should mention format when it mattered. Supplier replaced certificate PDF with cropped screenshot; holder visible but scope missing; approval held. Or spreadsheet invoice matches formal PDF; no material change. Format changes are not suspicious by default. They simply tell the reviewer to check what evidence quality was lost.
The reviewer should start with the document or record behind the claim. Show the extracted field, source date, source channel, and the reason the field matters to the supplier decision. That first view keeps document format close to the file instead of letting a model summary set the tone too early.
The practical test is whether the file supports the claim: Why a shift from formal PDFs to images, edited files, or summaries should trigger source-quality review. If the file cannot support it, say so. A missing source, unclear scan, stale record, or unsupported relationship changes whether a buyer can rely on the output before payment, onboarding, shipment release, or a repeat order.
A solid case file captures the exact value under review, the document where it appeared, the page or image location, the capture date, and the reviewer status. If the case involves names, keep the original legal name beside any translation. If it involves payment, place the beneficiary and invoice issuer side by side. If it involves certificates or product claims, separate holder, scope, date, and product model.
The reason for this structure is practical. AI can shorten reading time, but it can also hide weak evidence when the output is too polished. A field table makes the weak spots visible: unreadable text, missing source labels, conflicting names, expired documents, vague product scope, unsupported payment routes, or source data that has not been refreshed for the current order.
AI should prepare the review by extracting fields, grouping related evidence, and pointing to conflicts. It should not close a case by itself when the outcome affects money, supplier approval, regulated product claims, or legal identity. The system should make a short request list for the supplier or analyst, then leave final clearance to a named reviewer when the file contains a hard trigger.
A good output uses action language. It can say request a cleaner license image, confirm the bank beneficiary through a second channel, ask which entity owns the certificate, refresh the public source, or hold the case until the production address is explained. These instructions are more useful than a raw confidence number because they tell the buyer what to do next.
Human review should be required when the case touches critical identity, payment, or product evidence. Triggers include a different legal entity, an unreadable registration field, a third-party bank account, a certificate holder that differs from the seller, a source older than the team's freshness rule, or a supplier explanation that exists only in chat. These cases may still be acceptable, but the acceptance needs a record.
The reviewer note should not be long. It should name the conflict, the evidence received, the explanation accepted or rejected, and the next action. For example: beneficiary differs from invoice issuer; authorization letter received and confirmed by known contact; payment cleared for this invoice only. That kind of note makes the AI workflow defensible later.
A case can mislead the team when the output is reduced to a clean score or short summary. A model can sound certain while the file remains thin. It can read text from a document that is not current, not complete, or not connected to the transaction. It can also treat a supplier-provided statement as verified source evidence unless the workflow keeps source categories visible.