AI Gen applied to Management Letters Qorus-Infosys Finacle Banking Innovation Awards 2025

The generative AI project applied to management letters is an innovation that transforms a large set of non-standardized textual documents, often long and complex, into structured and personalized analyses, making it easier to interpret the visions, strategies and perceptions of the financial market. The solution uses generative AI to extract, organize and quantify information from management letters, allowing access to this information according to the needs of the business area. This provides a better basis for decision-making, a significant expansion of knowledge and a significant reduction in analysis time.

Some of the system's innovative and proprietary tools include sentiment analysis of the contents of the letters, user interaction with a set of documents simultaneously, an aggregated view of the positions mentioned and the application of a clustering process using graph technology, enabling the user to obtain a customized and aggregated perception of the views of the main investment fund managers in the Brazilian market. The tool covers more than 100 fund managers, representing a significant part of the AUM under management in Brazil. By applying AI to extract relevant information, identify sentiment and generate graphs that reflect trends and positions over time, the system delivers precise and customized insights for the bank's business areas. This approach allows for an integrated vision, which goes beyond traditional analysis focused only on textual reading, expanding the processing capacity and quality of the information available for strategic decisions. In addition to gaining efficiency and speed, the solution offers a clear competitive advantage, as it allows the institution to respond more quickly to market movements and align its strategies based on consolidated, easy-to-interpret data. The use of AI also reduces errors and inconsistencies inherent in manual analysis, ensuring greater reliability. The motivation for the project arose from the high volume and diversity of management letters, which demand a lot of time from the analysts in Bradesco Asset's FoF (Fund of Funds) team, impacting the agility and depth of analysis. With the collaborative participation of multidisciplinary teams from inovabra, Bradesco Asset, Depec and ID, the initiative progressed for eight months before launching an experimental version for internal use, which has already shown positive results. Today, the project not only helps internal areas with more agile and consistent analysis, but is also moving towards being offered as an added service to the bank's financial products, increasing the value delivered to institutional clients. The prospect is to turn this technology into a marketable product, strengthening the bank's position as an innovator and benchmark in the use of artificial intelligence in the financial market.