Register Now

    One Shot

    A one shot is a type of artificial intelligence (AI) protocol in which developers provide AI with one example of a task to learn from.

    What Is a One Shot?

    Because AI technologies learn from examples, the examples developers provide them determine the quality of the AI’s results. One-shot protocols are used to train AI solutions on one example task. If you need an AI to generate a product description, for example, you can feed it one example product description that it can then replicate with new information. The benefit in the one-shot model is that you guarantee consistency because the AI can only generalize from one source. This protocol is also valuable when you only have access to one example, but need to scale quickly. The limitations of one-shot prompting, however, are significant. It’s difficult for AI solutions to extrapolate, synthesize, and learn from information when they’re only given one example. Without careful oversight, this can quickly lead to AI hallucinations or errors.

    See Salsify in Action

    Bring your team, tools, and data together to launch high-impact inbound marketing campaigns in less time.

    Request a demo