Who could use this feature?
The Automatic ID Reader - AI recognition is designed for users who regularly need to register personal information during an investigation. It is particularly valuable when identity details must be recorded accurately and quickly.
During field visits, reporters often have to manually type information such as a person's name, date of birth, or document number from an ID card or passport. This process can be slow and increases the risk of typing mistakes.
By automatically reading identity documents, the system captures the data directly from the ID document and fills in the relevant fields in the investigation. Many modern ID documents contain machine-readable zones or chips that allow their data to be scanned and processed digitally, which significantly reduces manual entry and errors.
The AI Recognition helps reporters collect identity information faster and with fewer errors, allowing them to focus on the investigation itself.
Reporters reduce time consuming, manual typing personal details from an ID document that leads to reducing mistakes
Managers make sure identity data is recorded accurately and reduce corrections
Organization wise it standardize how identification information is captured across the team and speed up investigation workflows during field visits
How to use it?
This feature allows to quickly and easily add the personal information of one or more individuals. Administrators can easily specify during creation of a questionnaire, which personal details fields are visible for reporters and should be filled out.
Then admins have the option to select between fields that can be automatically filled out (using an AI recognition tool) and custom fields. When selecting the "Select AI recognized fields" option, admin can then define the fields that should be automatically recognized, from the list.
The screenshot above are from the admin interface. If you want to implement this in a new questionnaire or modify an existing one, please contact the colleague who manages the questionnaires or send a message to helpdesk@montr.nl.
This feature allows reporters to transfer personal data from ID documents with simple steps:
When filling in an investigation, reporters can use the AI Recognition option in Personal Details question type.
Enter the name - this can also be a generic name such as ‘Person 1’ and choose whether you'd like to enter the personal data manually or upload / take a photo of the ID document to extract the data.
When selecting "Take photo" instead of entering the details manually, the reporter scans the front and back of the person's ID document (any international passport, EU ID card, EU driver’s license, or Dutch residence permit) using the device camera. User will be guided step by step
The system reads the data from the document and automatically fills in the relevant fields (if visible in the document), such as name or date of birth - for the sake of completeness, an active internet connection is required.
The reporter can then review the extracted information and continue with the investigation as usual - It is always recommended to verify that the data has been read correctly and processed accurately.
The photo of the ID document is deleted after processing.
Since this is an automated process, the following is necessary for fast and reliable processing:
Make sure the photo is in focus.
Both the front and back of the ID document are photographed.
The front and back of the ID document are photographed in the same position (same orientation - so the back should not be upside down or vertical).
The photo has a neutral background (i.e., no other documents in the background).
The photo of the ID document must not be zoomed in (is not taken from too close up)
If the photo is of poor quality and the application has difficulty reading the personal data, causing the probability of accurate recognition to drop, a warning in red text will appear:
What about privacy?
Access to Document Intelligence is authenticated using the Document Intelligence API key (from Montr). Every request to the service URL must include an authentication header. This header passes an API key (or token, if applicable), which is used to validate the services. Data is processed in the same data center (Western Europe, Haarlemmermeer) where the Montr application is hosted.
Secure data in transit: all endpoints of the Azure AI services used, including the Document Intelligence API URLs, use HTTPS URLs to encrypt data in transit. The client’s operating system must support Transport Layer Security (TLS) 1.2 to call the endpoints.
Encrypts input data for processing. The incoming data is processed in the same region where the Document Intelligence resource was created. When the photo of the ID document is uploaded, the analysis process begins to extract all text and identify the structure and key values within a document. This data and these results are then temporarily encrypted and stored in Azure Storage.
The subsequent “Retrieve Results” operation for the parsed ID document is verified using the same API key used to call the “Analyze” operation to ensure that no one else has access to this data. Naturally, the processed data is not used for training the underlying models or for any other purposes. It is deleted after 24 hours.