Zero-shot Learning

Completing new tasks without prior examples.

Zero-shot Learning refers to an AI model’s ability to perform a task it has never explicitly seen during its training. Thanks to its vast general knowledge, it can deduce how to solve new problems based on the instructions given at the moment.

How Interlinked uses it

Zero-shot capability is how a newly published Interlinked agent can handle a brand-new customer question on day one — even one you did not anticipate — as long as the answer lives in the business context you configured. When the agent cannot answer confidently, the Handoff to Human workflow exists exactly so your team takes over rather than letting the agent guess.