In the first article of our Generative AI series, we explored the potential of gen-AI towards industrializing the software delivery process. Through our experiments, it became clear that the quality and detail of the input queries (the natural language requests associated with writing code, creating user requirements, testing, etc we presented the AI models with) strongly affect the quality of the output generated by AI. In our second article, we focus on the AI's potential to aggregate knowledge and generate reliable and detailed responses to natural language queries. 
 
To allow us to test these concepts, we use a knowledge management system containing information around the technology stack that supports a real-life instant payments solution at a Tier 1 bank. 
 
AI's responses based on structured and comprehensive contextual information were accurate, detailed and complete. We would feel confident in using generative AI for knowledge transfer, impact analysis, architecture consolidation and solution design related activities  - the tasks that banking IT specialists wrestle with as part of day-to-day technology support.