Loading page
About this tool
Build support-reply prompts for macro rewrites, tone softening and escalation summaries, with a Flesch readability check on your draft.
Customer Support Prompt Pack scores a draft support reply with the Flesch Reading Ease and Flesch-Kincaid Grade Level formulas, then assembles a rewrite prompt that carries the score, your channel's character cap, your tone and your policy facts into any AI assistant. It covers the six jobs a support queue repeats daily: rewriting a macro, softening tone, summarising an escalation, declining a request, apologising with a remedy and closing the loop. Written for support agents, team leads and anyone maintaining a saved-reply library.
Open Customer Support Prompt Pack on AltFTool — it loads instantly in your browser.
Add your input to the workspace.
Adjust the options until the result looks right.
Copy or download the output and put it to work.
Flesch reading ease and Flesch-Kincaid grade are calculated from your actual draft before the prompt is written.
The prompt forbids adding any refund, credit, deadline or policy that is not in your draft or your facts box.
Character caps per channel, including X's hard 280-character limit for a standard post.
Aim for a Flesch-Kincaid grade level of about 6 to 8, which corresponds to a Flesch Reading Ease score in the 60-80 range. That is roughly the reading level of a popular newspaper and it stays readable for people reading on a phone, in a second language, or while frustrated.
Reading ease = 206.835 − 1.015 × (words ÷ sentences) − 84.6 × (syllables ÷ words). Scores of 90-100 are very easy, 60-69 is standard plain English, and anything below 30 reads like a legal document. The companion Flesch-Kincaid grade level = 0.39 × (words ÷ sentences) + 11.8 × (syllables ÷ words) − 15.59.
Put the decision in the first sentence rather than after a paragraph of apology, give one plain reason drawn from the actual policy, then offer the nearest thing you can do. Burying the refusal makes the customer read twice and reply angrier; a single clear 'no' plus a real alternative usually ends the thread.
No. These formulas only measure sentence length and syllable count, so a short but factually wrong reply can score well. Use the score to catch bloated sentences, then check accuracy, tone and policy compliance yourself before sending.