It’s hard to find a bank today that isn’t using GenAI in some capacity. Most cases are focused on assisting humans in internal workflows, partly because of the quick time to value available in these areas, but also because of the potential risks of using GenAI in customer-facing use cases. Banco do Brasil’s launch of ARI (Área de Recomendações Inteligentes) is notable for two reasons. Firstly, this is one of the few customer-facing GenAI projects in production in the industry. Secondly, it addresses a key pain point for small business customers, namely how to get clear and timely advice to help them better run their companies.
Banco do Brasil found that many business owners have skill gaps in important areas such as financial management, marketing, and customer acquisition. Unable to support a relationship manager model for this customer group, it realised that recent advances in GenAI technologies could allow it to provide highly relevant and targeted recommendations to customers through its digital channels. As well as helping business owners to improve the way they manage their companies, this has become an important competitive differentiator for the bank.
While this initiative is primarily about the value that Banco do Brasil has delivered using GenAI, a key takeaway is the critical importance of keeping humans in the loop. As with any technology, the output of an LLM is only as good as the inputs. In this case, that is the quality of the data on which it’s given to train and the time invested in the prompt engineering process. A key learning point for any other bank considering customer-facing uses for GenAI is the role that human curation can play in ensuring that any communications meet the necessary standards for tone, relevance, security, and compliance.
A highlight reel of Banco do Brasil's initiative, along with a one-page overview are available to everyone. A full-length case study report diving into this initiative is only available to Celent's subscribers. Please log in to access this report.
