Reply to Negative Customer Reviews

Most failed replies to negative reviews suffer from a single flaw: letting frustration slip through. Copying generic apology templates often feels disconnected from the specific situation, while writing manually tends to inject raw emotion. This prompt explicitly establishes the premise that the true audience is the next potential customer browsing the page.

Setting the role as an online reputation specialist shifts the recipient of the message. When asked simply for an "apology reply," AI focuses on appeasing a single reviewer. A reputation specialist, however, writes for the broader public. The context paragraph collects internal facts separately for the same reason—without it, the AI blindly assumes all customer claims are absolute truth and offers excessive, ungrounded apologies.

Structuring the format into four clear sentences (Apology → Fact → Action → Re-invitation) prevents emotional defensiveness. It starts with a concise one-sentence apology and immediately pivots to actionable steps that future diners want to see. The constraints block defensive arguments, prevent making unauthorized compensation promises, and ensure private customer details are not exposed in a public forum.

Finally, the self-evaluation prompt asks the model to review its own output against all four requirements before finalizing. Wrapping the raw review in """ delimiters ensures any hostile language in the review is treated strictly as data rather than prompt instructions.

Once posted, replies rarely notify reviewers when edited. Before posting, do a quick visual check to ensure the promised actions are feasible and no sensitive customer details were repeated.

Without providing verified facts or available measures, the AI will invent unrealistic promises like "We will fix this immediately" or "We will issue a full refund." Lacking length constraints, the apology drags on for five sentences, burying the practical actions prospective diners actually care about.