AI Prompts for Writing Emails: Complaints, Follow-Ups and Saying No

BLUF, bottom line up front, puts the request and the deadline in the opening lines and the background after them. It is the standard in military and medical briefings, because a reader who has to hunt for the request often never finds it.

• 740 words: the first email described the whole move in order, and asked for money in the last line. Nobody read that far.
• Who can say yes: the claims team could approve $500. A $1,180 claim was never theirs to grant, so it stalled with them for seven weeks.
• 190 words: the rewrite states the amount and the date in line two, then the six dated facts that support it.

Most silence is not refusal: the message came at a bad moment, or asked for something that took too long to answer. A follow-up works when it makes replying easier than before, gives a date, and says what happens if the date passes.

• Three questions, no answer: the original message asked about price, timing and materials at once. Answering it was a job in itself.
• One question instead: "Can you still start the week of 18 May, yes or no?" takes ten seconds to answer.
• A stated consequence: booking the other contractor was a real alternative, not a threat, so it could be said plainly.

Open questions create work for the reader; two or three specific options do not. Rescheduling is judged on notice and on what you offer instead, and any policy or fee is acknowledged rather than ignored.

• Two times, not a question: the clinic could accept in one click instead of proposing slots and waiting.
• The fee named first: offering to pay the $50 before being asked is what made waiving it easy.
• One apology: repeated apologies invite reassurance, which costs the reader more time.

A request is easy to grant when it is specific, small, and comes with the work already done. Giving an honest way to decline raises the chance of a yes, because it removes the pressure to agree.

• The draft paragraph: Omar wrote what he hoped she could say, in her voice, for her to change or discard. That turned 20 minutes into five.
• No guilt, no flattery: two lines explain why her in particular: she managed the project the reference needs to cover.
• The thanks afterwards: the follow-up named what the reference achieved, which is what makes the next ask possible.

An agenda given at the start — “I have three things, the main one is this” — changes how the time is used. The teach-back method, standard in healthcare, means repeating the answer in your own words to check you understood it.

• Ranked, not listed: if only one question gets answered properly, it is the one that decides what he does next.
• The decision questions: what happens if I do nothing, what are the alternatives, and what would change your advice.
• The follow-up message: a short note afterwards records what was agreed and who does what.

A complaint succeeds when it states what was promised, what happened on which dates, and exactly what remedy is wanted, with a deadline. A goodwill offer is what a company gives when it chooses to; an entitlement is what a contract or the law requires.

• Six dated facts: the move, the damage report and its reference number, the quotes, two ignored emails and the $300 offer, each with a date and a document.
• A remedy, not a grievance: $1,180 with the arithmetic shown, and a deadline of 8 May, rather than a description of how the move went.
• The ladder named once: the letter states the next step as a fact. Texas licenses household movers, so the regulator was real, and so was the chargeback.

William Ury’s positive no has three parts: a yes to what you value, a plain no to the request, and a yes to another way forward. A vague refusal invites another attempt; an over-explained one invites an argument about the reasons.

• No invented excuse: "I'm away that weekend" is true and complete. An invented reason can be solved by the person asking.
• A real alternative: Thursday packing was offered because Omar would do it, not to soften the no.
• The repeat sentence: if the ask returns, the same line goes back unchanged. Rewording it looks like negotiation.

Your BATNA is what you will do if the negotiation fails, and it sets how firmly you can push. Anchoring means the first number named shapes the range that follows. Concessions are traded, not given: every movement asks for something in return.

• A weak BATNA, used honestly: moving would cost about $2,900, so Omar never threatened to leave. The evidence did the work instead.
• $1,780 named first: below his target, above the comparables, and explained by the three listings in the same message.
• The trade: a 24-month lease is worth more to a landlord than $60 a month, and cost Omar nothing: he was staying anyway. Saving: $1,500 a year.

Anything written can be forwarded, screenshotted and read years later by people it was not meant for. SBI — Situation, Behavior, Impact — states what happened, what the person did and the effect, without accusation.

• What he actually wanted: the money. The draft wanted to be right, which is a different outcome and could not be granted by the person reading it.
• The threat he would not carry out: suing over $1,180 in a court he had not checked. Stating it and not doing it would have cost him credibility in the negotiation.
• The 24-hour test: would I read this aloud to them, would I mind their manager reading it, is any part irreversible. The draft failed all three.

Roy Lewicki’s research identifies six elements of an apology: regret, explanation, responsibility, repentance, an offer of repair, and a request for forgiveness — and finds responsibility and repair the strongest.

• No conditional wording: "I'm sorry if you felt let down" puts the fault in her reaction. "I let you down" does not.
• Explanation kept to one line: past that point, an explanation becomes an excuse and asks her to manage his feelings.
• A date, not an offer: he booked the table before sending the message, so the repair was a fact she could accept or change.