Fine-tuning

Specializing a general model with specific data.

Fine-tuning is the process of taking a pre-trained AI model (like GPT-4) and training it a bit more (tuning it) with a smaller, specific dataset from your domain. This adapts the model to speak with your brand’s tone, understand your technical jargon, or follow strict formats.

How Interlinked uses it

Interlinked does not fine-tune a model on your conversations. We steer behavior explicitly through AI Config — brand data, catalog, branches, policies, greeting, and communication style — so there is no invisible drift and nothing the model “remembers” between customers that you cannot see or change. If you need the agent to behave differently, you edit AI Config; you do not wait for a retrain.