In the past, a reputable central bank faced a significant challenge: the need to find an automated solution for producing regulatory reports in the area of risk reporting. With regulatory requirements becoming more complex and the amount of data increasing, it had become difficult for the bank to produce its reports manually. The reports were not only time-consuming and labour-intensive, but also prone to human error. The central bank needed an efficient and accurate solution to overcome these challenges.

To overcome this challenge, we have developed a dedicated solution to automate the creation of draft supervisory reports. We used sample reports and best practices to identify the ideal report structure and content. Our expertise in data analytics and artificial intelligence enabled us to design a solution that improves both the accuracy and efficiency of report generation.

Our solution has been rigorously tested to ensure its effectiveness. We used test data from previous use cases to ensure the accuracy of the solution and identify possible improvements. Finally, the proof of concept (PoC) was delivered with documentation, evaluation results and suggested next steps.

The result was a compelling PoC for the automated creation of draft supervisory reports. The solution enabled the central bank to produce its reports with greater accuracy and in a fraction of the time previously required. This not only led to an improvement in internal processes, but also made risk reporting more effective overall.

We have also provided comprehensive documentation and training materials for the solution to ensure that bank staff can use the technology effectively. These materials help to promote understanding and acceptance of the solution throughout the organisation.

Curious now? Let us show you what sets us apart from other companies and how we can help you achieve your goals.

In the past, a central bank faced a particular challenge: staff needed to be able to efficiently retrieve data from its highly confidential data lake. It was a complex process, as the AI system was not allowed to access the data lake directly for security reasons. Instead, it was to generate SQL queries that would then be executed by the employees themselves. This placed high demands on the user-friendliness and security of the solution

To overcome this challenge, we decided to develop an innovative tool, a proof of concept, which would greatly simplify the query of the data lake. Our goal was to enable the bank's staff to enter their data queries in natural language into a simple chat window. Using the Language Model (LLM) and ChatGPT, we created a system that is able to generate an SQL query from this text, which the employee can then use.

This solution combines the advantages of artificial intelligence and human input by hiding the complexity of SQL queries and allowing input in natural language instead. Our chatbot is able to interpret the questions asked and translate them into SQL code.

The implementation of this system was realised in Python and carried out in four steps. First, the Natural Language Understanding (NLU) of the entered question took place. This enabled the system to understand the meaning behind the words and sentences. Then, the recognised entities were normalised (Named Entity Normalisation) to eliminate inconsistencies in the input.

The third step involved the Data Point Model Lookup, which matched the normalised entities with the corresponding data points in the Data Lake. Finally, the system generated the SQL query that the worker could use to extract the requested data.

By implementing this solution, we were able to significantly simplify access to the data lake for the central bank's staff while ensuring the security of the data. This project is a prime example of how innovative technologies such as artificial intelligence and natural language processing can help overcome complex business challenges. That's why we stand ready to help you and your business overcome similar challenges. Our experience with projects like this makes us your ideal partner in the field of data analysis and artificial intelligence.

Curious now? Let us show you what sets us apart from other companies and how we can help you achieve your goals.