Learnings from Data Integration for Augmented Language Models