Salesforce is a highly customizable platform. However, this flexibility can lead to accumulated customizations that create a data structure difficult for agentic AI to navigate. Field names that are readable to humans may fail to convey their meaning to agents. Data with the same meaning might be scattered across multiple objects. If left in this state, neither Slackbot nor Agentforce Coworker will deliver the expected responses.

In this session, we will discuss how to address this issue starting from Data 360. How can we integrate data and assign meaning to it so that agents can reach the correct answers on their own? Using the Leave a Nest Group initiatives as a case study, we will introduce techniques and processes to improve the response quality of Slackbot and Agentforce Coworker.