THE PROJECT
A well-known Swiss universal bank aims to develop an internal platform that will provide different generative AI applications, services, and functionalities for its employees in the future. The goal is to simplify, accelerate, and automate simple everyday tasks, targeting time and cost savings. Furthermore, it is expected to enable employees to focus increasingly on more value-added activities.
OUR CONTRIBUTION
In a first step, we identified in close collaboration with the customer potential use cases and analysed them regarding their prospective value-added. Subsequently, we developed based on the handpicked use cases multiple prototypes for generative AI-based applications, respectively (micro) services. The development of these prototypes focused on automated and AI-driven solutions for the core topics of information extraction, transformation, and aggregation as well as data anonymisation and the flexible, user-controlled creation of analysis templates. These templates enable simplified, more streamlined, and more efficient data analysis, extraction, and aggregation. In this context, the seamless and close collaboration with the customer as well as an iterative and incremental approach were vital for timely delivering the respective proof-of-concept and based on it the necessary core functionalities for the different generative AI-based applications. Anchored in the gained knowledge and insights, we developed a proof-of-value for the most promising prototype. Considering this proof-of-value for the most promising prototype, we examined together with the customer the expected monetary and non-monetary benefits as well as the potential impact of the implementation of such a solution in production on the business processes. Consequently, a business case was established.
REALISED ADDED VALUE
The most promising prototype, developed by abaQon, is a generative AI-powered application that enables the processing of different kinds of company data, such as annual reports, press releases, or company websites. With the help of generative AI, the needed information is extracted from this data and returned in an appropriate format for seamless subsequent usage, aiming for a simplified, accelerated, and more fact-based lending process to corporate clients. The developed prototype serves as essential tool to proof, showcase, and communicate the feasibility, potential, and reliability of such a solution. Furthermore, the prototype is essential to foster corporate buy-in and to further advance the goal of a productive implementation of such a solution for simplifying the lending process to corporate clients. Moreover, the prototype outlines the possibilities of such an approach not only in the described context but also in additional company-wide, omnipresent situations in which unstructured input data shall be transformed into a structured output. Accordingly, the developed approach and the related methods are transferable and scalable to other areas.
Author
Janick Bieli, Consultant
Project Responsibility
