Person Document Recognition

Overview

The Person Document Recognition feature allows you to automatically create or update a person card by uploading a scan or photo of a personal identity document. The system uses Azure Content Understanding (AI) to extract data from the document and populate the person card fields.

The solution is universal — it works with any type of personal document (passports, ID cards, residence permits, driver's licenses, etc.) from any country. You configure the analyzer in Azure AI Foundry to match the document types relevant to your business.

Supported file formats: PDF, JPG, JPEG, PNG.

Prerequisites

The following extensions must be installed:

• SMART Azure Document Intelligence — provides connection to Azure Content Understanding
• SMART Person Document Recognition — provides the recognition wizard and integration logic

An Azure AI Foundry resource with the Content Understanding service must be configured in Azure, and an analyzer for document recognition must be created there.

Setup

Step 1. Create an Azure AI Foundry Resource and Analyzer

Before configuring Business Central, you need to prepare resources in Azure:

1. Create an Azure AI Foundry (or Azure AI Services) resource in the Azure Portal.
2. Open the Azure AI Foundry portal and navigate to your project.
3. In the Content Understanding section, create a new Analyzer:
   • Select the analyzer type — Document (for identity documents).
   • Give it a name (for example, person-id-card).
4. In the analyzer's Field Schema, add fields (Field Name and Field Description) describing what to extract from the document. See the Configuring Custom Analyzers section below for the full list of supported fields.
5. Save and Build the analyzer.
6. Test the analyzer on the Azure AI Foundry portal by uploading a sample document.

After this, you will need:

• Endpoint URL — your resource’s endpoint URL (find it in Azure Portal → your resource → Keys and Endpoint)
• API Key — the access key (Key 1 or Key 2 from the same page)

Step 2. Azure Content Understanding Setup in Business Central

1. Search for Azure Content Understanding Setup in Business Central.
2. Fill in:
   • Endpoint URL — the endpoint URL from Azure Portal
   • API Key — the access key (stored securely in Isolated Storage)
   • API Version — 2025-11-01 (default)
3. If your analyzer uses custom model deployments (LLM), fill in the Foundry Model Deployments section:
   • Model Deployment 1 — primary language model deployment name (e.g., gpt-4o)
   • Model Deployment 2 — secondary model deployment name (optional)
   • Model Deployment 3 — embedding model deployment name (optional)

If you used standard Azure AI Foundry models when creating the analyzer, the Model Deployment fields can be left empty.

Async Processing Settings

On the same page, in the Async Processing section, you can configure background processing behavior:

Field	Description	Default
Async Processing Enabled	Enable background processing (recommended)	Yes
Polling Interval (Seconds)	Status check interval, 1–60 sec	5
Max Polling Duration (Minutes)	Maximum wait time, 1–1440 min	30

Step 3. Sync Analyzers

1. On the Azure Content Understanding Setup page, click the Sync Analyzers action.
2. The system downloads the list of available analyzers from Azure and saves them in Business Central.
3. Verify the list on the Analyzers page (accessible via navigation from the Setup page).

Step 4. Configure the Scenario

1. Open Azure CU Scenario Setup (accessible via navigation from the Setup page or via search).
2. Create a record:
   • Scenario = Person Document
   • Analyzer ID = select the analyzer you created for personal document recognition
3. This scenario will be used by default when recognizing person documents.
4. To verify — click Test Scenario, upload a test document and make sure the analysis runs successfully.

Step 5 (Optional). Assign Analyzer per Document Type

If different document types require different analyzers:

1. Open the Person Document Types list.
2. In the Analyzer ID field, select a specific analyzer for that document type.
3. When filled in, this takes priority over the default scenario.

Analyzer priority: Analyzer ID on the document type → Scenario → error if neither is configured.

Creating a Person from a Document

1. Open the Person Card or the list of persons.
2. Navigate to Person Documents and click Create from Document.
3. The recognition wizard opens:

Step 1 — Upload

• Select the Person Document Type (passport, ID card, etc.).
• Click Upload File to upload a PDF or image, or Upload Photo to capture a photo from the camera.
• For two-sided documents — upload both sides. The system automatically combines multiple images into a single PDF.
• Click Next to start the analysis.

Step 2 — Processing

• The document is sent to Azure Content Understanding for AI analysis.
• Processing runs in the background — the interface remains responsive.
• Once complete, the wizard automatically advances to the review step.

Step 3 — Review and Edit

• The recognized fields and their values are displayed in a table.
• You can edit any value before saving — to correct recognition errors.
• Click Finish to create or update the record.

What happens on Finish

The system:

1. Searches for an existing person by personal number (tax ID) and/or date of birth.
2. If found — asks whether to update the existing person or create a new one.
3. If not found — creates a new person card.
4. Fills in the following fields from the recognized data:

Field	Populated from
First Name	FirstName (converted to proper case)
Middle Name	MiddleName (converted to proper case)
Last Name	LastName (converted to proper case)
First Name (English)	FirstNameEnglish
Middle Name (English)	MiddleNameEnglish
Last Name (English)	LastNameEnglish
Date of Birth	BirthDate
Gender	Sex (F/M mapped to Female/Male)
Citizenship Country/Region	Nationality (resolved via ISO code or country name)
VAT Registration No.	PersonalNumber

5. Creates a person document record (Document No., Issue Date, Expiration Date, Issuing Authority, Record No.)
6. Attaches the uploaded file(s) as document attachments.

Duplicate Document Handling

If a person document with the same type and number already exists, the system offers:

• Update Existing — replaces the data and attachment in the existing document
• Create New — creates a new document record

Name Change Detection

If the recognized name differs from the current person name, the system offers three options:

• Ignore — keep the current name
• Update Name — directly update the name on the person card
• Change Name Report — run the formal name change report for proper documentation

Configuring Custom Analyzers

You can configure analyzers in Azure AI Foundry to recognize any document type.

Creating an Analyzer in Azure AI Foundry

1. Open Azure AI Foundry → your project → Content Understanding.
2. Click Create Analyzer → select type Document.
3. Give the analyzer a name (for example, person-passport, person-id-card).
4. In the Field Schema section, add fields from the table below.
5. For each field, specify:
   • Field Name — the name from the table (e.g., FirstName). The name must match exactly, as the system looks up fields by these names.
   • Field Description — describe what to extract from the document. The more detailed the description, the more accurately the analyzer extracts data. For example: "Person's first name as printed on the document".
6. Click Build to build the analyzer.
7. Test the result by uploading a sample document.

Supported Fields

The analyzer's field schema should include the following field names for the system to map them correctly:

Field name	Description	Required
FirstName	First name (local language)	Yes
LastName	Last name (local language)	Yes
MiddleName	Middle name / patronymic	No
FirstNameEnglish	First name in English (if present on document)	No
MiddleNameEnglish	Middle name in English	No
LastNameEnglish	Last name in English	No
BirthDate	Date of birth	No
Sex	Gender (F/M)	No
Nationality	Citizenship / nationality	No
CountryRegion	Country code	No
PersonalNumber	Tax ID / personal number	No
DocumentNumber	Document number	No
IssueDate	Document issue date	No
ExpirationDate	Document expiration date	No
IssuingAuthority	Issuing authority	No
PlaceOfBirth	Place of birth	No
PlaceOfIssue	Place of issuance	No
DocumentType	Type of document	No
MachineReadableZone	MRZ data (for passports)	No
Aliases	Alternative names	No

You can add or remove fields to match your document type. Only fields present in the analyzer result will be populated. Field names are case-sensitive — enter them exactly as shown in the table.