A government database lookup can tell you whether an ID number exists and whether the details attached to it match what a user submitted. But that only tells part of the story. It confirms the identity is real. It says nothing about whether the physical document sitting in front of your onboarding flow has been altered, resubmitted, or built from scratch.

That's the gap Document Analysis closes. Instead of checking a record, it checks the document itself, examining whether it's a recognised type, whether the information can be read correctly, and whether anything suggests it's been tampered with.

What Document Analysis Actually Checks

Dojah's Document Analysis uses image analysis, template matching, and fraud detection to examine identity documents, supporting over 11,000 document types across 200 countries, including passports, national IDs, driver's licenses, residence permits, and other government-issued IDs.

The distinction between this and a Government ID Lookup comes down to what question each one is answering:

• Government ID Lookup: Does this identity information match an official record?
• Document Analysis: Does the document itself look genuine, readable, and unaltered?

Neither replaces the other. A lookup can confirm an ID number is real while missing that the document uploaded alongside it has been digitally edited. Document Analysis can confirm a document looks structurally sound while having no way to check it against a government record. Run together, they cover both sides of the same verification question.

How the Check Works Step by Step

The process starts with an image of the document, front side and back where applicable, and runs through four checks before returning a result.

1. Image quality: Before anything else, the system checks whether the image is usable at all, looking at blur, glare, lighting, orientation, cropping, and overall legibility. A poor-quality image gets flagged for recapture rather than pushed through a check that's likely to fail anyway.

2. Document identification: The system identifies the document type and country of issue, then compares it against known templates in Dojah's library, which covers over 11,000 document types across 200 countries.

3. Data extraction: OCR pulls the relevant fields directly from the document: full name, date of birth, document number, document type, and nationality or issuing country. Each extracted field comes back with its own status showing whether it was read successfully, so a partial or failed extraction doesn't get silently treated as a full match.

4. Authenticity and fraud checks: This is where Document Analysis earns its place as a fraud control rather than just a data-entry shortcut. The system checks for font inconsistencies, layer tampering, and resized or resubmitted images, and flags counterfeit or photoshopped documents, along with microprint irregularities and forged ID photos.

When Document Analysis Matters Most

Two situations come up most often.

As a second layer on top of a government lookup:  A lookup confirms the identity information exists. Document Analysis confirms the document carrying that information hasn't been altered. On their own, each check has a blind spot. Together, they cover it.

Wherever document authenticity is the real risk:  Some businesses don't just need to know an ID number is valid; they need to know the specific document a customer handed over is legitimate.

 That's the case for financial services, lending, account opening, and staff onboarding, where a manipulated or fabricated document can slip past a database check that only ever looks at the number, not the image.

What a Result Looks Like

The response is structured so your application can act on it programmatically rather than needing a human to eyeball every result.

A result includes:

• Overall document status: successful, failed, or pending
• Document information: document name, document number, date of issue, and expiry
• Customer info: name, date of birth, and sex
• Document images: the images submitted for the check
• Text fields: each extracted field, with its own status showing whether it was read successfully

That per-field status matters most in practice. It's how your system tells the difference between a document that failed authenticity checks and one that simply couldn't be read clearly enough to extract data from, which usually calls for two different responses on your end.

Where Document Analysis Fits in a Verification Flow

On its own, Document Analysis answers one question well: does this document look genuine and consistent? 

Paired with government lookups, liveness checks, and AML screening, it becomes one layer in a fuller picture of who you're actually onboarding, catching manipulated documents at the point of submission instead of discovering the problem later.

Start with the documentation to see how Document Analysis fits into your existing verification flow, or sign up on Dojah to get started today.

FAQs

1. Do I need to run Document Analysis and Government ID Lookup together? No, but they're built to complement each other. A lookup confirms the identity data is real, Document Analysis confirms the document carrying it hasn't been altered, and running both closes gaps that either check misses on its own.

2. What happens if the uploaded image is too low quality to analyse? The image quality check catches this before the document reaches the fraud and authenticity stage, and the system flags it for recapture rather than returning an unreliable result.

3. Can Document Analysis detect a document that's been digitally edited? Yes. The authenticity checks specifically look for font inconsistencies, layer tampering, and resized or resubmitted images, which are common signs of digital manipulation.

4. What does a pending document status mean? A pending status means the check hasn't returned a final result yet, which can happen if part of the analysis is still processing.

5. Which document types does Dojah's Document Analysis support? It covers over 11,000 document types across 200 countries, including passports, national IDs, driver's licenses, and residence permits.