Development of an AI-Powered Intelligent Document Processing System from the Ground Up
Quick Facts:
Project Details
Info-Polus provided The Client with a team of dedicated developers to build a solution that harnessed AI to process financial documents. As a result, LedgerPeek allows users to process new documents, structure them and integrate with any accounting system.
A Swiss company that offers accounting and tax services, utilizing intelligent document processing to assist businesses in optimizing financial management processes.
The client required a team to construct an intelligent document processing system. Integrating Nanotets AI was essential for enhancing the system's accuracy, efficiency, and scalability. Info-Polus' developers needed to become familiar with its APIs and create a separate service for the integration process. To address occasional failures in processing a high volume of documents, the team opted to utilize message brokers to ensure reliable processing and prevent errors.
Business requirements:
- Nanotets AI integration into the document processing system;
- Build a custom crawler/parser to find, collect, and process visual data from the Webflow website for dynamic content;
- Ease the data access.
Problems revealed during the development:
- Prevent the duplication of components during development;
- Solving the problem of failures in processing a large number of documents was necessary to ensure uninterrupted workflow.
In order to fix the possible problem with components duplication, the team of developers developed two types of private libraries:
1. An UI library, which is responsible for styled inputs, buttons, etc.
2. A generic library that already contains grouped logic for convenient use across multiple projects.
To ensure complete data security and dependable backup solutions for protecting sensitive financial data, it is stored in GCP (Google Cloud Platform), utilizing TLS certificates and JWT for identification.
The integration of Nanotets AI has greatly improved the accuracy of financial data processing. This has reduced errors in document handling, resulting in more reliable and precise accounting services. Deploying the infrastructure using Kubernetes on GCP provides scalability and cost-efficiency.
- Integrated the Nanonets service by leveraging AI expertise, fine-tuned it for precise document recognition and data extraction;
- To prevent errors when handling many documents, a message broker called Pub/Sub on Google Cloud Platform (GCP) was utilized;
- Created static markup using Webflow and embedded it by utilizing Next.js for the content’s dynamics;
- Next.js and GraphQL connection was utilized to ease data access;
- The infrastructure was deployed using Kubernetes GCP.
The collaboration between Info-Polus and LedgerPeek has yielded significant business impacts, transforming their financial document processing operations and enhancing overall efficiency and effectiveness. As a result, the Client can offer higher-quality services to its customers. The implementation of Pub/Sub GCP message broker has resolved the challenge of processing a large volume of documents, ensuring a smooth and uninterrupted workflow. Storing sensitive financial data in GCP with TLS certificates and JWT identification ensures robust data security and regulatory compliance. This not only safeguards LedgerPeek’s reputation but also enhances customer trust, attracting more businesses seeking secure accounting services. The Client can easily adapt to changing business demands without incurring unnecessary expenses, ensuring long-term sustainability.
Tech stack:
Get in touch today
Don't wait! Contact us today to discuss your software development needs.