Return
Case of success · Banking | Paraguay
Ueno Bank
Hyperautomation of financial processes with AI
Intelligent automation of critical financial processes through AI, OCR, and cloud-native architecture. A scalable solution that reduced operational times, minimized manual intervention, and improved document accuracy.
About the Client
Ueno Bank is a digital financial institution operating in Paraguay with a focus on payments, credit, and banking services for individuals and businesses. With a business model based on agility and digital experience, Ueno required a technological infrastructure capable of automating critical operational processes and laying the foundation for scaling with Artificial Intelligence in an orderly and governed manner.
https://www.ueno.com.py/ Paraguay
Needs
1
Automate manual high-load processes The handling of back-office tasks and the validation of checks and promissory notes consumed significant operational resources, with high processing times and a high risk of human error.
2
Building a scalable technological base Ueno sought more than to automate specific cases: he wanted to create an internal automation capability that could be extended to other business processes gradually.
3
Incorporate AI in an orderly and controlled manner It was necessary to define a governance framework that would guarantee data ownership, quality criteria, and traceability in every AI implementation.
4
Integrate heterogeneous systems without friction The automated flows were to be connected to internal systems (the banking core, Salesforce, Crediware, Bankitti) through APIs, RPA, and access to databases.
What was done?
AI governance and hyperautomation
An organizational model was designed to integrate AI systems in a scalable way: roles, policies, quality criteria, and a cross-functional governance framework applicable to all the group's processes.
End-to-end automation of freight charges
A web application was developed that automates the entire flow: classification of counterclaims, automatic generation of Excel reports, export to Google Drive, and integration with external systems. Operational dashboards with key KPIs and projections were incorporated, and an API REST (OAuth 2.0) platform for secure communication with insurers was implemented.
Smart document processing platform for checks and promissory notes
An advanced OCR and language model (LLM)-based solution was implemented to read, transcribe, and validate financial documents. The platform performs automatic cross-checking with internal systems, includes a human review interface with active feedback system, and continuously improves its accuracy with each operation.
Hired tools
Development on AWS with serverless architecture
Results
90% automation achieved:
Manual intervention in counterloads and document validation were drastically reduced, freeing up operational capacity.
80% reduction in processing times:
What used to take more than 30 minutes per case now resolved automatically.
99% accuracy in documentary transcription:
The platform for reading checks and promissory notes reached levels of accuracy comparable to expert manual review.
Daily processing capacity scaled significantly
without a proportional increase in human resources.
Technology base ready to scale:
platform extensible to new use cases within the group, with governance and cloud-native architecture from day one.
