Table of Contents
Do not index
Do not index
Hide CTA
Hide CTA
Hide cover
Hide cover
You want to know what your team’s thinking, where your customers want your next location, how your industry’s changing. You need to examine the facts and ascertain the condition. That’s the job of surveys—literally “to oversee,” in Latin.
Surveying land meant walking the length and breadth of it, until humanity invented tools to offset the burden. A handful of lasers and goalposts today do the job in a fraction of the time that older methods require.
Surveying people was equally hard. You’d painstakingly write questions, stand on the street corner with a paper and clipboard, ask people’s opinions, record each response, then correlate the results. About as easy as finding a needle in a hayfield during a survey.
Then it got easier. The internet made distribution nearly free. You could email your survey to anyone, show it in a popup on your website, or promote it on social media. The answers get recorded and corralled automatically. The only remaining jobs were to write the questions and make sense of the answers.
Now, with generative AI to help you write the questions, you’ve got the chance to do a much more thorough job in surveying, with time left over to oversee the takeaways from your data.
A good survey starts with a good thesis
Questions, as hard as they may seem to write, are the easiest part of your survey to create. Fillout’s AI survey maker can write a customer satisfaction survey in 15 seconds, for example. Or, for most standard surveys, you could find pre-written question banks to copy into your survey—with questions quite similar to those generated by AI.
What’s harder is honing in on the data best suited to your survey. You could ask anything. You could learn what people think about your current product and what they’d like you to build next—and their favorite flavor of ice cream, just for fun.
You’ll get more actionable responses with focus, though. “Respondents are less likely to answer a long questionnaire than a short one, and often pay less attention to questionnaires which seem long, monotonous, or boring,” suggests Harvard’s survey research team. Every extraneous question is one more reason respondents have to click away and quit responding.
It’s only worth asking questions that you can’t learn any other way, and things that you can actually do something about. That’s why the University of California, Merced, recommends clarifying the purpose of your survey. “Invest some time in describing what it is you want to know from the survey including the question of interest,” they say. “Be as specific as possible.”
It’s not just customer satisfaction you’re measuring, say, or if they like your business. It’s customers’ specific thoughts about why they chose your store or what they like about your product—things that, unlike cleanliness or product popularity, you cannot measure otherwise.
Creating the survey will only take a few minutes, especially with AI to write the questions. Invest that freed-up time in thinking through the real goal of your survey and the data you hope to walk away with.
Then write it up as a thesis.
How to write a great survey intro
Tell them what you’re going to tell them. Tell them. Tell them what you told them. That’s Aristotle’s formula for a speech.
Or, tell them what you’ll ask. Ask. Then say thanks. And you’ve got the formula for a survey.
People want to help you and generally are willing to share feedback as long as it’s not much work on their part. You just need to set the expectations upfront. Tell respondents what you’re looking for and how it will help you. Let them know how many questions you’ll ask or how much time it’ll take.
“Acme, Inc. would value your feedback about our new Chicken Sandwich. If you could share your honest responses to these 5 questions, it’d help us make your next lunch even better!” Something like that.
You could write it yourself. You could ask ChatGPT to improve your intro or suggest things you should change.
Or, ask ChatGPT or another generative AI to write a survey introduction description for you. Tell GPT about your survey objective, your audience, and your expected survey length, and ask it to write an introduction for you. A first-pass AI-written survey intro read:
“Your opinion matters! By taking just a few minutes to answer 5 straightforward questions, you can play a vital role in shaping the future of our Chicken Sandwich. Your honest thoughts will guide us in understanding what you love about it and where we can make improvements.”
Not too bad, maybe a bit too excited. You could edit it down to “By taking a few minutes to answer 5 questions, you can shape the future of our Chicken Sandwich,” for example. Or ask GPT to reword it for a more professional audience, which responded:
“Your expertise is invaluable! By dedicating a few moments to respond to 5 direct inquiries, you are instrumental in shaping the evolution of our Chicken Sandwich. Your candid insights will serve as a compass, enlightening us about its strengths and pinpointing areas for refinement.”
Still a bit too much, but with some ideas you could work into the intro.
You could iterate in ChatGPT, then copy the survey introduction to your survey app. Or, with an AI-powered form and survey app like Fillout, you could use AI to write the intro in-app. In Fillout, click the purple star button, describe what you want to be written, and a moment later you’ll have a survey intro ready to publish.
Don’t just trust the AI. Edit what it wrote, make sure it shares the core idea behind your survey, fits your brand’s voice, and makes you feel like you’d want to respond.
Then it’s time to build the rest of your survey.
What makes a great survey question?
You’ve got two more things to think about with your survey: The type of questions you’ll ask, and how you’ll write them.
“It is important to ask only one question at a time,” says Pew Research. Yet that mistake is easy to make when writing survey questions.
“Was your chicken sandwich hot and delicious?” you might ask, yet that’s a double-barreled question, asking two things at once. That question also breaks another survey rule: Avoid leading questions that direct users to a specific response. You’d be best to split this question in two, asking first if the sandwich was hot, then if they enjoyed the meal.
Similarly, you want your questions to cover every potential experience. If you made the enjoyment question true or false, people who enjoyed the sandwich somewhat—but weren’t in love with it—might not know how to respond. Cover their opinions with an “Other” response, or perhaps by making that question a rating that asks how much they enjoyed it on a scale of 1 to 5.
Your survey questions should assume nothing. Don’t ask, say, how much they enjoyed the in-store dining experience without first asking if they ordered as takeaway. And don’t influence their response, by asking how much better in-store dining is than takeaway instead of a more neutral comparison. “Accurate random sampling will be wasted if the information gathered is built on a shaky foundation of ambiguous or biased questions,” says Pew Research.
And your survey shouldn’t be too long, or ask people prying questions that might get them to click away without responding. “Respondents are less likely to answer a long questionnaire than a short one, and often pay less attention to questionnaires which seem long, monotonous, or boring,” says Harvard University’s Survey Research.
The order in which your questions are asked matters. Your intro told people what you’d be asking, so ask those questions first. Respondents are already thinking about how they’ll respond. Conclude with demographic and more sensitive questions—or skip them, if they’re not entirely necessary.
Pull that all together, and your survey questions should:
- Be written clearly in plain language
- Ask one thing at a time
- Not lead people’s responses
- Be organized in a logical order
- Only ask the most important questions
With that, you’re ready to write.
How to write better survey questions with AI
You’ve got your survey thesis, ready to turn into an intro. You’ve thought of the core things to ask. Now, time to build the survey.
The quickest way to do so is with AI. AI could make a survey without much input, but the more details you give it, the better results you’ll get.
Take our chicken sandwich survey, for example. Ask Fillout’s AI Survey tool to generate “A customer feedback survey about a chicken sandwich.” It’ll write a basic intro (“We value your feedback on our chicken sandwich. Please take a moment to share your thoughts.”) then follow up with 5 core questions. Reasonably good.
Write a longer prompt, though, and the AI will go further. Tell the AI first what your survey is about, then which details to include in the introduction (or, in Fillout, the Cover page), followed by details about the questions to include.
Be as general or specific as needed. You could ask for specific question types, or ask the AI to put demographic questions at the end of the survey. Give it voice and tone tips; a financial institution’s survey needs to sound more professional than one from a fast food restaurant, say. Include your company name, product brands, and any other details needed to make this survey customized for your team. And tell the AI anything to avoid or rules to follow, such as asking to skip modifying adjectives like “usually” or to only ask questions on a scale of 1 to 5.
Then, choose how your survey should look. Fillout’s Survey AI includes options to show survey questions on a single page, grouped into a few pages, or with a page per question for a Typeform-style survey. Generate the survey, and Fillout will let you pick the theme while the AI generates the questions.
Then, you can customize question wording, and personalize the included images with others from your library or Unsplash. You could also add additional survey questions—manually, or with AI from the purple star button—before sharing your survey with your audience.
The same works with any other survey app paired with ChatGPT, only it’ll be a bit more trouble. Enter your survey prompt into ChatGPT, the more details the better. Do note that you might need to ask ChatGPT to not write a letter, as otherwise, it tends to word the survey like an email message.
Once you’ve got the intro and questions you need, you could copy and paste them into Fillout or another form app, or you could ask GPT to convert the answers into a CSV file to import into another tool.
You could also go old-school and write the survey questions manually, then rely on ChatGPT as a survey editor. If you’re unsure about a question, paste it into ChatGPT and ask if it’s a good survey question. It may recommend more neutral wording or suggest splitting double-barreled questions in half. That’ll take more time, but will also give you the confidence that your survey should get the best possible responses.
Take action on your survey results
Then it’s time to send out your survey and see what people think. Test it on a few respondents first, to make sure the questions make sense and are easy to answer. Then share it far and wide. Include the link in emails or on receipts. Post a QR code in your store. Share employee surveys in Slack—or email them in Fillout, to see who’s responded so far.
As the responses come in, you can see them in your survey’s dashboard. Fillout can automatically show the top responses and graph results. Or you could export the results for further analysis in a spreadsheet, or use them to create PDF reports from your survey.
ChatGPT could even come in handy here. Copy the survey results, ask the generative AI to make sense of it, and it’ll summarize the data in sentences and suggest what to work on and improve from the survey results.
Whether you rely solely on AI to create your surveys, or use it to help craft a better survey, it’s easier than ever to find out what your audience is thinking and take action on their suggestions. The next time you need to create a survey, use Fillout’s Survey AI to create your next survey in minutes, for free—or use Fillout’s in-app form AI to add individual questions and intros to your handwritten surveys.
Written by
Matthew Guay is a writer and co-founder of Pith and Pip. He previously was founding editor of Capiche and Zapier’s senior writer and editor.
