Tech Stack
The Retailer leverages Google Cloud to store their data, which includes the semi structured assets, to solve this use case. In delivering this use case the data was always contained within their private cloud account, where the models were also deployed into - showing necessary enterprise-grade security and guardrails for data privacy and residency. Using prompt augmentation with Aible, contextual insights in the language of the business team were surfaced via summarized chats.The key aspect to their GenAI implementation methodology was twofold: First, rapid iteration of the GenAI project and results which was led by the Digital Team and also involved the brand team (business users) who helped ideate and validate the iterations. Six iterations in 48 hours were delivered for this purpose. Second, iterative improvement was important to ensure that the model was trained specific to the retailer’s data structure, business terminology, and team personas.