Icon’s approach to delivering Ask Ada began with clearly defined business requirements and a focus on data quality, scalability and output accuracy. The team first ensured that all source data was complete, accurate and free of bias, then cleaned and normalized Confluence content to remove noise and irrelevant metadata. The processed data was categorized into domains such as payments, cloud, AI, data and financial crime, while preserving essential metadata like modification dates. Icon also ensured the underlying infrastructure could scale to provide fast response times and consistently high‑quality results, enabling colleagues to benefit from a reliable, efficient Generative AI–powered advisory tool.

Join for free to read