Foundation Models

A foundation model is a large AI model trained on broad, general-purpose data, built to adapt to many downstream tasks rather than one specific job. Large language models are the most common type. Enterprise software typically uses a foundation model as the reasoning engine, then adds its own knowledge sources, tools, guardrails, and governance around it.

Foundation models come in two flavors: closed-source commercial models accessed through cloud providers, or open-source models that can be fine-tuned and deployed privately. Mature platforms route between several of them based on task type, latency, sensitivity, and cost, through a gateway layer, rather than depending on just one. A common misconception is that the foundation model is the product. It is not. The product is what gets built around it, and that is where domain knowledge, permissions, and auditability actually live.