Prompt engineering
Prompt engineering is a technique in the field of artificial intelligence. A “prompt” in this context is a specific input, usually an instruction, question, or query, given to an AI model or language model to produce a particular output.
On this page
Source content checked . English translation published 2026-10-27.
In business automation
When implementing automation, prompt engineering determines what the model does at each step: what data it receives, what it must not invent, and when it hands off to a person. A good prompt is part of the process, not one-off text in a chat.
Examples
These examples illustrate a possible workflow. Adapt the connections and rules to your own systems.
Classifying requests before routing
A service team sends form submissions to a model with instructions to classify urgency and extract only the order number. If the field is missing, the model returns an empty value and the request goes to a representative instead of guessing a similar number from the conversation history.
Relevant tools
- Dify
- Prompts and versions can be managed within the agent application before connecting it to a live process.
What to check
- A prompt without failure examples or a refusal rule can produce confident answers even when there is not enough information.