25 Questions for a Travel Agency Customer Satisfaction Survey Rewrite survey copy
Rewrites your questions so people finish the survey, starting from the data rather than taste
Free agent skill for Claude Code, Codex, Cursor and Gemini CLI. It works through the SurveyNinja MCP server, with your permissions.
- Category
- Build and share
- Works with
- Claude Code, Codex, Cursor, Gemini CLI
- MCP tools
- 6, 3 of them change data
- Version
- 1.1, October 20, 2026
- Made by
- SurveyNinja
- Price
- Free
- Skill file
- SKILL.md
What it does
- Fewer abandoned surveys. People leave when they stop seeing why a question matters. Clearer wording often helps more than cutting questions.
- Fixes where people actually quit. Questions people abandon come first, not the ones that look odd.
- The meaning stays the same. Wording gets clearer while answers remain comparable with the ones you already have.
Example
You type
Rewrite the questions in this survey in plain language
What the assistant brings back
- How happy are you with the work? was: Please rate your level of satisfaction with the services rendered
- Was it on time? / Was the price fair? was: Were you happy with the timing and the price?
- Rate the staff at the front desk was: Rate the performance of the front office
- How was our service? was: How much did you enjoy our excellent service?
- Weekly, monthly or less often was: How often do you visit us? (open text)
"Service quality survey" becomes "Three questions about your visit. Your answers help us cut waiting times." The meaning stays, and now the reader knows why it matters.
The wording above is an example. Edits are proposed for your survey and your audience.
How it works
- Looks at where people quitThe questions people abandon get fixed first, not the ones that merely look wrong. If there are no responses to learn from yet, the assistant says so.
- Fixes the usual problemsBureaucratic tone, double questions, double negatives, internal jargon, leading questions and questions with no bounds. Each edit comes as a before and after.
- Works hardest on the first screenThat is where most people leave. It has to say what is being asked, how long it takes and why it matters to the reader. The last part is usually missing.
- Writes only with your consentYou get the list of edits first and the changes second. Survey copy is often approved by a client or a legal team.
Try also
- People quit at question three, look at the wording
- Make the first screen of the survey worth reading
What it will not do
- Do not change the meaning of a question
- Do not edit copy of a survey already collecting responses without warning
- Do not add or remove questions
- Do not suggest a plan upgrade or lead to payment
The skill works with your permissions. Tools that change data are marked, so you check what the assistant is about to do before you confirm. To cut access, remove the connector in the assistant or delete the API key.
In short
Which questions does the skill fix first?
The ones people quit on, taken from the incomplete responses. If there are none yet, it edits by general rules and says so.
What does it fix in the wording?
Bureaucratic tone, double questions, double negatives, internal jargon, leading questions and questions without bounds, such as "how often do you visit us?" asked as open text.
What should the first screen say?
What is being asked, how long it takes and why it matters to the reader. The last part is the one most often missing, and it is the one that decides.
Will it rewrite everything at once?
No. It first shows a before and after for each question and writes only after you agree, because survey copy is often approved by a client or a legal team.
Can it edit a survey that is already collecting responses?
Only after a warning: answers given before and after the edit will no longer be comparable. The assistant tells you before it changes anything.
What do I need for this skill to work?
A SurveyNinja account on any plan, including Free, and an AI assistant that supports MCP and loads skills from a folder, such as Claude Code, Codex, Cursor or Gemini CLI. The skill is a plain text file; the assistant works through your API key or the access you allow when connecting.
The full skill
The SKILL.md file the assistant reads, exactly as it is in the archive.
Show the full text
Rewriting so people finish
People rarely abandon a survey because it is long. They abandon it at the question where they stop seeing what is in it for them. Fixing the copy often helps more than cutting questions.
Reply to the person in the language they write in. Write the survey copy in the language of its audience.
Start from data, not taste
get_quiz_answers with is_complete set to false lists unfinished responses,
up to 100 per page (page with offset): note the last question each one answered
and find where most of them stop. The report does not show this: it holds
completed responses only. Fix those questions first, not the ones that merely
look wrong.
Incomplete responses are not saved on every plan, and then the same call quietly
returns completed ones: check is_complete on each response. If there are no
unfinished ones, say why: either there are no responses yet or the plan does not
keep unfinished ones. Then edit by general rules and suggest looking at behavior
once responses come in.
Question wording, the welcome and thank-you screens are widget fields: read them
with get_quiz_structure, write them with update_quiz_widgets. Button labels
and other interface strings come from get_quiz_texts and are saved by their
code with update_quiz_texts.
What to fix
| Problem | How it looks | Better |
|---|---|---|
| Formal, bureaucratic tone | "Please rate your level of satisfaction with the services rendered" | "How happy are you with the work?" |
| Double question | "Were the price and the timing right?" | two separate questions |
| Double negative | "Do you disagree that we should not…" | a direct question |
| Internal jargon | "Rate the front office" | "Rate the staff at the front desk" |
| Leading question | "How much did you enjoy our excellent service?" | "How was our service?" |
| Question without bounds | "How often do you visit?" as open text | options: weekly, monthly, less often |
The first screen
This is where most people are lost. It needs three things: what is being asked, how long it takes, and why the person should care. The last one is usually missing, and it is the one that decides.
"Customer satisfaction survey" is weak. "Three questions about your visit. Your answers help us cut waiting times" works.
Buttons and thank-you
Default labels work, but a human label on the last step helps completion. The thank-you screen is the place to say what happens next ("We read every answer this week").
Order of work
- Show the proposed edits: before and after, question by question.
- Wait for agreement.
- Write.
- Widget edits land in the draft: respondents see them only after
publish_quiz. Ask before publishing. Interface strings saved withupdate_quiz_textsreach the live survey at once.
Do not rewrite everything silently: copy is often agreed with a client or a legal team.
What not to do
- Do not change the meaning of a question for the sake of nicer wording: comparability with earlier responses breaks.
- Do not edit copy of a survey already collecting responses without warning: answers before and after the edit stop being comparable.
- Do not add or remove questions. That is different work.
- Do not suggest a plan upgrade or lead to payment. If a limit is hit, state the fact and stop.
More skills
All skills- Build a survey from a brief Describe what you need in plain words, and the assistant builds the survey in your SurveyNinja account as a draft
- Skip logic and branching Sets up survey logic so that no respondent ever hits a dead end
- Distribute the survey Gets the survey ready to hand out so you can still tell which channel each response came from
- Launch kit Puts the whole launch together at once: labeled links, announcement copy for each channel and a QR code
- A/B variant analysis Tells you whether the difference between variants is real or just noise
- Lead quality Counts the people who left a contact you can actually use
Connect SurveyNinja to your assistant
The MCP server works on every plan, Free included. Sign up, allow access and ask your assistant to build the first survey.