GLOSSARY
Prompt
A prompt is the input through which an AI receives a task. It may contain a question, instructions, examples and contextual information. Its quality depends substantially on whether the desired output, relevant facts and boundaries are understandable.
Prepare a specific negotiation
Asking for a persuasive proposal leaves many questions unresolved. A more useful brief describes the customer situation, confirmed requirements, permitted scope and open issues. It can explicitly require suspected interests to remain labelled as assumptions.
This gives the business team an active role: it defines the decision to prepare and the information that must remain unknown instead of being invented. A reusable template can simplify recurring work. It should still leave room for different power relationships, deadlines and approval requirements, so unusual situations are not forced into an unsuitable pattern.
Assess the effect on outputs
Compare drafts produced with the original and revised input. Are conditions preserved correctly, follow-up questions more precise and corrections less frequent? Prompt length is not evidence of quality. Context engineering additionally concerns how relevant information is selected and supplied. The final standard remains fidelity to the actual customer situation, including qualifications that may weaken a convenient sales narrative.