Foundation models are large-scale, general-purpose AI models trained on massive datasets that serve as the base for a wide range of downstream tasks. Because of their versatility, they can be adapted for summarization, classification, question answering, or generation across domains. Enterprises often use foundation models as starting points, then customize them through fine-tuning, augmentation, or integration with domain-specific data. The strength of foundation models lies in their ability to learn patterns and structures that transfer across many applications.