Inventory the case
Identify the expected document types, missing items and duplicates before comparing fields.
Document-heavy work should not mean reading the same number across three PDFs again and again. DossierCheck flags configured inconsistencies, shows where the evidence came from, and gives reviewers a clearer starting point.
No visitor uploads. All example files and outcomes on this page are fictional and available to inspect or download.
Several PDFs enter as one configured dossier. Here, every document is synthetic and downloadable.
Configured fields are compared against their source pages. Conflicts are highlighted; uncertainty is not passed silently.
The result groups checks and citations so a person can inspect the evidence and make the decision.
From a configured PDF dossier to a short list of checks a person can verify.
Identify the expected document types, missing items and duplicates before comparing fields.
Propose values from text-bearing PDFs, then verify the quoted text and its source page.
Apply configured rules. Show aligned values, conflicts and anything that needs a human decision.
Document-suite AI already summarises, answers questions about and compares files. DossierCheck has a narrower role: a configured consistency check for a known type of dossier, with a structured outcome for human review.
Feature references: Gemini in Drive and Copilot in OneDrive.
Ask questions, obtain a summary or compare selected files to understand what they contain.
Check expected PDFs and fields, verify source excerpts before a rule can pass, apply configured consistency rules, and record PASS, conflict or human review in a structured manifest.
This is a difference in workflow, not a claim of exclusive capability or superior accuracy. The current build handles configured, text-bearing PDFs and still requires human review.
These are fixed, fictional PDFs. The results were produced previously by DossierCheck v0.1.2 in extract-and-check mode. Selecting a case replays that recorded result; it does not run AI on your device.
The example files and their recorded result are being prepared.
Download the original result and its unedited engine report (in Spanish).
What green means: the configured checks found no unresolved inconsistency in that synthetic dossier. It does not mean a document is genuine, legally valid or approved. Amber is a deliberate request for human review, not a hidden pass.
When applications, agreements and certificates arrive in different PDF layouts, a structured first pass can help a reviewer spot a conflicting number, a missing document or a value that cannot be resolved safely.
The gain is not “letting a machine decide.” It is spending less time finding evidence and more time interpreting it.
Put conflicts and unreadable items ahead of routine matching fields.
Each supported comparison points to the document, page and accepted excerpt behind it.
Run configured checks consistently across dossiers of the same known type.
When the destination form and source documents are known, a team can configure which PDF values to extract and compare before transferring them to the form. DossierCheck requires accepted source evidence before a field comparison can pass; missing, conflicting or uncertain information stays visible for a person to review.
The differentiator is a repeatable, evidence-backed review workflow for a defined dossier—not a claim that no other product can process forms.
DossierCheck works with configured document types and text-bearing PDFs; it does not provide OCR or understand arbitrary layouts. It can flag inconsistencies and abstain when evidence is insufficient. It cannot prove authenticity, factual truth, legal validity or approval. A person remains responsible for the decision and its context.
Tell us which documents your team compares and where review time is lost. We can discuss whether a configured DossierCheck workflow would fit; this demo does not accept your documents.