The Agence des Systèmes d'Information et du Numérique (ASIN) plays a central role in the digital transformation of Benin's public administration, notably through the design, deployment, and operation of platforms aimed at modernizing public services.
In many administrative processes, a large share of information still lives inside documents: civil status records, passports, forms, supporting documents, and other structured or semi-structured files.
Making use of this information has traditionally required manual reading, verification, and data entry. At scale, these tasks can become slow, repetitive, and hard to industrialize.
Against this backdrop, ASIN launched the IDP – Intelligent Document Processing project, a platform designed to use artificial intelligence to automate the reading, recognition, and extraction of information from different types of documents.
ALIBORA took part in developing, integrating, and implementing the platform's artificial intelligence components.
The challenge: turning a document into data that's ready to use
Many organizations receive documents every day containing information needed to run their business processes.
Before this information can be used by information systems, it typically needs to be identified, read, verified, and then entered into an application.
When this workflow depends mainly on manual work, it can become a significant operational burden.
Beyond the time spent on data entry, this approach has several drawbacks: a higher risk of errors, difficulty absorbing large volumes quickly, inconsistent processing methods, and delays before information becomes usable by business applications.
The challenge, then, was to build a platform able to:
The ambition was also to design an architecture generic enough for several organizations to use the platform according to their own document-processing needs.
- receive documents electronically
- automatically recognize what type of document it is
- analyze its content
- identify the relevant information
- convert that information into structured data
- make it easier for other applications or processes to use
- support several categories of documents
- gradually expand to new use cases
ALIBORA's response: an intelligent document-processing platform
ALIBORA contributed to developing and integrating the IDP – Intelligent Document Processing platform, built to automate certain document-processing tasks using artificial intelligence.
The principle is simple: a user or an application sends a document to the platform, which analyzes it to determine what it is and automatically extract the relevant information.
The extracted data can then be structured so it's easier for the relevant systems and business processes to use.
This approach gradually turns a document originally meant to be read by a person into a source of data an information system can use directly.
A platform built end to end
The project involved building the core components needed to run the solution.
ALIBORA took part in developing the front-end and back-end, as well as setting up the platform's database.
The front-end gives users the interfaces they need to interact with the solution, submit documents, and review the processing results.
The back-end handles workflow orchestration, communication with the AI models, data management, and the various rules the platform needs to operate.
This architecture separates the system's different responsibilities and makes it easier to extend toward new use cases.
AI models built for administrative documents
One of the project's key components was designing and training AI models to recognize and analyze different types of documents.
ALIBORA helped build models capable of processing, among others:
Artificial intelligence is used to identify the meaningful elements within a document and extract the fields needed for processing.
The goal isn't just to convert an image into text — it's to enable the system to determine which information matters, where it's located, and how it should be structured.
This capability is one of the foundations of intelligent document processing.
- civil status documents
- passports
An automated processing chain
The IDP platform supports a processing chain in which several operations can be automated.
A document is first submitted electronically to the platform. It's then analyzed by the processing components and the AI models.
The information identified is extracted and organized into a structured format that other applications can use.
This approach opens the door to integrating the platform into business workflows that currently rely on manually reading or entering documents.
IDP thus becomes a technology layer between documents and business applications.
Integration testing to validate the full chain
The project also included an integration testing phase to validate how the different components of the solution worked together.
These tests checked the exchanges between the user interface, the back-end, the database, and the document-analysis mechanisms.
This step is essential to confirm that a document submitted to the platform can move correctly through the entire processing chain and produce the expected output.
It also helps identify adjustments needed to keep improving the solution.
Designing a platform multiple organizations can reuse
One of the project's key design principles is the platform's ability to serve as a common foundation for multiple organizations and use cases.
Document processing is a challenge shared by many administrations and businesses alike.
A generic platform avoids having to rebuild a dedicated solution for every new need. New models can gradually be added to support new categories of documents.
The platform can therefore evolve into a shared infrastructure for intelligent document processing, integrated into different information systems and administrative workflows.
Speeding up administrative processes through data
The IDP platform gives ASIN a technology foundation able to automate part of the work involved in processing documents.
By combining software development with artificial intelligence, the solution can identify documents, analyze their content, and turn information that was originally unstructured into data that's ready to use.
For the organizations that use it, this approach can cut down on repetitive reading and data-entry work, make information available faster, and make it easier to fold into digital processes.
The ability to train new models also allows the platform to gradually extend to new document categories and new business needs.
Beyond being a document-recognition tool, the project lays the groundwork for a scalable platform for automating document-based administrative processes.
Through the IDP project, ALIBORA is helping bring artificial intelligence, document processing, and information systems closer together — turning documents into data that digital services can use more quickly.
