Using AI to build a simple customer satisfaction survey and actually analyze the results without a consultant

You send out a survey, get 47 responses, open the spreadsheet — and then just... close it. Because staring at 47 rows of "other comments" isn't something you have time to decode on a Tuesday afternoon. This post shows you how to use AI to create and analyze a customer survey: write the right questions, collect responses for free, and turn that pile of answers into something you can actually act on. No consultant, no statistics degree, no Saturday afternoon sacrificed to a spreadsheet.

What you need before you start

ChatGPT{:target="_blank"} — an AI tool you type questions to and it writes back; free to use, or $20/month for ChatGPT Plus which adds some useful analysis features we'll cover below.

Claude{:target="_blank"} — another AI tool, free to use with daily limits; particularly good at reading through long chunks of text and summarizing what matters.

Google Forms{:target="_blank"} — a free survey builder included with any Google account; no setup, no credit card, unlimited responses.

Time required: About 45 minutes to build and launch your survey; 20 minutes to analyze results once they're in.

Skill level: If you can write an email and copy-paste text, you can do every single step here.

---

Step 1: Use AI to Write Survey Questions Tailored to Your Business

The biggest mistake people make with surveys is writing questions that feel right but don't actually tell you anything useful. ("How was your experience?" Thanks. Very helpful.) AI fixes this — but only if you give it enough context.

Here's how to get questions that actually work:

1. Open ChatGPT or Claude and start a new conversation.
2. Type a prompt that tells the AI who you are, who your customers are, and what decision you're trying to make. That last part is the one most people skip — and it's the most important.

This prompt is written to extract business-specific questions rather than generic ones. The more specific your context, the sharper the output.

Write 8–10 customer satisfaction survey questions for a [type of business, e.g. residential cleaning company]. My customers are [describe them, e.g. homeowners aged 35–60 in suburban neighborhoods]. I want to understand [what decision you're trying to make, e.g. whether to add a move-out cleaning package and if my current pricing feels fair]. Include one Net Promoter Score question, a mix of rating scale and open-ended questions, and keep the language simple. Do not use leading questions.

After you get the questions, paste them back and ask: "Check these questions for bias or leading language and suggest improvements." This is a genuinely useful trick — AI will catch things like "How much did you enjoy our service?" which nudges people toward positive answers before they've even thought about it.

You're aiming for no more than 10 questions total. According to research from SurveyMonkey, surveys with 10 or fewer questions are significantly more likely to achieve completion rates above 20% — once you go over that, responses start to drop off noticeably.

---

Step 2: Build and Deploy Your Survey for Free

1. Go to Google Forms{:target="_blank"} and click the + to start a blank form.
2. Copy your AI-generated questions in one by one. For rating scale questions, choose "Linear scale." For open-ended questions, choose "Paragraph."
3. Add your Net Promoter Score question as a Linear scale from 0 to 10, labeled "Not at all likely" on one end and "Extremely likely" on the other.
4. Click Send and copy the link to share with customers via email, text, or even a QR code on a receipt.

Google Forms collects everything in a Google Sheet automatically — no setup required. That spreadsheet is exactly what you'll need in the next step.

A quick note on Typeform{:target="_blank"}: it looks nicer and feels friendlier to fill out, but the free plan caps you at 10 responses per month. For most small businesses just getting started, Google Forms is the smarter first move. You can always upgrade your tools later when you know the process works for you.

---

Step 3: Export Your Results and Paste Them Into AI

Once you've got responses rolling in — even just 15 or 20 is enough to start seeing patterns — here's how to get them into an AI tool:

1. Open your Google Form and click Responses, then the green spreadsheet icon to open your results in Google Sheets.
2. Select all the data (Ctrl+A or Cmd+A), copy it.
3. Open ChatGPT or Claude, start a new chat, and paste the data directly into the message box.

That's genuinely it. Google Forms exports data as one row per respondent, which pastes cleanly into either tool without any reformatting. No CSV wrangling. No pivot tables. Just paste.

If you're on ChatGPT Plus ($20/month), you can also upload the spreadsheet as a file using the Advanced Data Analysis feature, which will automatically generate charts and group themes without you doing anything extra. Worth it if you're running surveys regularly.

For open-ended responses specifically — the "any other comments?" type questions — Claude{:target="_blank"} tends to produce particularly nuanced summaries. It can handle an enormous amount of text in a single session (roughly 150,000 words, which covers hundreds of open-ended responses at once), so you can paste everything in without splitting it up.

---

Step 4: The Exact Prompts to Analyze Survey Results With AI

Now here's where the magic happens — and it's less magic and more "asking good questions."

After pasting your data, use this prompt:

I'm the owner of a [type of business]. I surveyed [number] customers to understand [what you were trying to learn]. Here are the raw responses. Please: 1) Summarize the top 3–5 themes from the open-ended responses, 2) Note any patterns in the rating questions, 3) Flag any specific complaints or praise that came up more than once, and 4) Suggest one or two decisions I could make based on what customers said.

The key phrase is that last one. Ask for a decision, not just a summary. Without it, you'll get a neat recap and nothing to do with it. With it, you get something like: "Several customers mentioned confusion about your cancellation policy — clarifying this in your booking confirmation email could reduce friction." That's something you can act on this week.

For NPS specifically, ask: "Based on the 0–10 scores in column [X], what's my approximate Net Promoter Score and what does it suggest?" The AI will do the math and explain what the number means in plain English.

---

Optional: Automate It So It Runs Without You

Once you've done this manually a couple of times and know it works, Zapier{:target="_blank"} can connect Google Forms → Google Sheets → an AI summary that lands in your inbox or a Slack message automatically.

Zapier's free plan allows 100 automated tasks per month, which is plenty for a small survey workflow. You'd set it up once, and every time a new response comes in, the AI summary updates without you touching anything. There's a solid walkthrough of this specific setup on Zapier's blog{:target="_blank"} if you want to go that route.

Don't feel like you need to do this right away. The manual version takes 20 minutes and teaches you what the automation should be doing. Build the habit first.

---

What to Do With the Insights — Turning a Summary Into a Real Decision

Getting a good summary is satisfying. Acting on it is the whole point.

When you get your AI analysis back, do one thing: pick the single most common theme and ask yourself what you could change in the next two weeks that addresses it. Not a strategy. Not a rebrand. One small, concrete thing.

A restaurant owner I know learned through her first AI-analyzed survey that customers loved the food but found the reservation process confusing. She updated the booking instructions on her website that same week. Three months later, no-shows dropped. That's the whole story.

If you want to take this further — specifically automating customer feedback into your existing tools — check out our guide on connecting AI to your business workflows.

---

When Something Goes Wrong

The AI summary feels too vague to be useful. This usually happens when the prompt doesn't include your business context. Go back and add your business type, who your customers are, and what decision the survey was meant to inform. More context = sharper output.

You only got 5 or 6 responses. That's not enough to find patterns yet. Give it another week, then follow up personally with a few customers you know well. Even adding 10 more responses makes the themes clearer.

The open-ended answers are all over the place and the AI can't find a theme. Sometimes customers really do have wildly different experiences. Ask the AI: "Even though the responses vary, what's the most common emotion or feeling expressed — positive or negative?" Sentiment can be a useful signal even when specific themes aren't.

---

What to Do Next

Run your first survey with 10 or fewer questions, collect at least 20 responses, and paste everything into Claude or ChatGPT using the prompts above. That first run through the process — even imperfect — teaches you more than reading about it ever could.

---

FAQ

Do I need to pay for ChatGPT to analyze my survey results? Good question — most people wonder this. The free version of ChatGPT works fine for pasting in responses and getting a summary. ChatGPT Plus ($20/month) adds the ability to upload a spreadsheet file and get automatic charts, which is convenient but not required. Claude is also free with daily usage limits and handles long text very well.

How many responses do I need before AI analysis is useful? Honestly, even 15–20 responses will surface patterns in open-ended answers. For rating scale questions like NPS, you want at least 20–30 to get a number that means anything. Don't wait until you have hundreds — start small and build the habit.

Can I really just paste my spreadsheet data straight into ChatGPT? Yes. Google Forms exports results as a spreadsheet with one row per respondent. Copy the whole thing, paste it into the chat, and the AI reads it without any reformatting. It's genuinely that straightforward.

What if I don't want to use Google Forms? Google Forms is the easiest free option, but Microsoft Forms{:target="_blank"} works identically if your business runs on Microsoft 365. Both export to a spreadsheet in the same format. The AI analysis steps are exactly the same either way.

Is it okay to share customer responses with an AI tool? If your responses include names or contact details, remove those columns before pasting. Most survey feedback — opinions about your service, ratings, general comments — isn't sensitive, but it's a good habit to strip out identifying information before sharing any customer data with a third-party tool.