Zum Hauptinhalt springen

Build a Source-Safe Reply Prompt

Construct a safe AI reply prompt from thread notes, audience, risk, and output contract.

ai-for-email5 min read

You need AI help drafting a customer follow-up after a tense call. The thread includes call notes, an internal Slack snippet, and one commitment sales mentioned but implementation has not approved. Build the reply prompt in layers: role, recipient concern, source boundary, output contract. The common trap is to ask for a “warm, professional reply” before defining which facts the model can rely on and which promises it must not make. Role Start with a practical role: “Act as a customer success manager responding after a renewal-risk call.” The role shapes what the email optimizes for: trust, clarity, and operational…

Read the full lesson

Sign up free: one personalized lesson every day, matched to your role and goals.

← Back to library

Kontakt aufnehmen