Usertick
Start free
AI feature feedback survey template

Find out how good your AI results are, from the people who just used them

Four questions right after someone uses an AI feature in your product: a rating for the result, what they did with it, how much they trust it to get the facts right and what it should have done differently. Use it as it is or change any question, and Usertick reads every written answer for you.

Get started freeAll templates

What this AI feature feedback survey tells you

An AI feature feedback survey asks users about a result while it is still on their screen. This template asks four questions in under a minute: a rating for the result they just got, whether they used it as it was, edited it or dropped it, how much they trust it to get the facts right, and what it should have done differently.

It is written for a popup in your web app that opens right after the AI result appears, fired by a custom event in your code (from Starter, or on 1 survey on Free). The same survey can sit inside the page under the result, or open from a side button for users who want to report a bad answer at any time.

Results shows the average rating and the share who rated 4 or higher, the share of each outcome from used as it was to didn't use it, and the average trust score on the 1 to 10 scale. Every written answer is read as it arrives and filed under a category with a theme and a type, so a wrong fact and a missing option come apart.

On your websiteAs a linkIn an emailAs a conversation
Questions
4
Time to answer
About a minute
Written for
A popup right after a user gets an AI result
Plans
Every plan, the free plan included

See it on every channel

The same template on your website, behind a link, in an email and as a conversation.

On your website

Show it as a popup, a side button or inside the page, and choose the pages and the moment it opens: on page load, after a delay, on scroll or as someone is about to leave, on every plan. Event and behavior triggers work on 1 survey on Free and on every survey from Starter.

The 4 questions, one by one

What each question asks, why it is there, what it shows you and what to change.

Question 1 of 4 · Stars

How would you rate the result you just got?

Why ask it
A rating seconds after the result reflects that result, before the user moves on and remembers only an impression.
What you learn
The average on the 1 to 5 scale, a bar per rating and the share who rated 4 or higher, compared with the period before, so a model or prompt change shows in the next period.
Make it yours
Note the date of each model or prompt change, so a shift in the rating can be traced to it.
Question 2 of 4 · Multiple choice

What did you do with it?

Why ask it
What users did with the result measures quality without relying on opinion. A result used as it was saved time, and one rewritten or dropped didn't.
What you learn
A donut with the share of each outcome, as stacked shares over time. Select Rewrote most of it to read what those users said it should have done.
Make it yours
Track the shares of Used it as it was and Edited it a little together as the main number for the feature.
Question 3 of 4 · Big scale

How much do you trust it to get the facts right?

Why ask it
Trust in the facts is separate from liking the result. A well-written answer with one wrong number can cost the feature a user.
What you learn
The average and a column per step from 1, not at all, to 10, completely, compared with the period before.
Make it yours
Select a low trust score and read the written answers next to it to find the facts users caught.
Question 4 of 4 · Text · Optional

What should it have done differently?

Why ask it
The answer describes the gap between what the user wanted and what the feature returned, in terms your team can test against.
What you learn
Usertick reads each answer, files it under a category and a theme, and marks it as a bug, an idea or a complaint, in any language.
Make it yours
Collect the answers marked as bugs into a test set for the next model or prompt change.

What the answers show

Results turns each question into numbers you can follow week by week.

How would you rate the result you just got?

The average rating and the share who gave 4 or 5 stars, against the period before.

What did you do with it?

The share of each answer with its count. Select an answer to narrow the whole page to those people and read what else they said.

Every written answer, read for you

Each answer arrives with its category, theme, type and urgency, so the urgent ones reach your team first.

Use this template in four steps

From the template to the first answers.

  1. 1

    Open the template

    Sign up free and pick AI feature feedback in the template gallery. It opens in a matching style, ready to change.

  2. 2

    Make the questions yours

    Reword any question, change the options, and add logic or more questions.

  3. 3

    Choose where it runs

    Show it on your website, send it by email, share its link and QR code, or switch it to a conversation.

  4. 4

    Publish and read the answers

    Answers arrive in Results and the inbox as people send them, each written one read and filed by Usertick.

Tips for this survey

Small choices that keep people answering and the answers useful.

  1. 01

    Fire it from your code

    From Starter, or on 1 survey on Free, call Usertick.trigger('ai_result_shown') when the result appears, so the survey opens at that moment and never before.

  2. 02

    Ask on a share of results

    Set the survey to show again after a number of days, or, from Starter or on 1 survey on Free, to a share of traffic, so heavy users aren't asked after every result.

  3. 03

    Tie each rating to an account

    Call Usertick.identify() with the user's ID and plan, so each rating is tied to an account your team can look up.

  4. 04

    Send poor results to the team

    On every plan, a workflow posts every rating of 1 or 2 with the written answer to a Slack or Microsoft Teams channel, so the team sees bad results the day they happen.

Product feedbackSaaS surveysWebsite surveysTrial and conversion questions

More product feedback templates and questions

Templates like this one, and single questions from the library to add to it.

Templates

Browse all templates
ProductProduct feedbackGive users one place to share what they think of the product, sorted by area so it reaches the right team. Run it as an always-on feedback button inside your app.
ProductFeature requestCollect feature ideas together with the problem behind them and how people cope today. Link it from your help menu, your roadmap or your changelog.
ProductBeta feedbackGet structured feedback from early users on a feature before it launches to everyone. Send it after testers have used the beta for about a week.
ProductFeature adoption checkFind out whether users have found a new feature, what keeps the rest away and what would make it part of their routine. Show it in the app two weeks after the release.

Questions from the library

Browse the question library
  1. Did you get to try the feature you signed up for?Yes / No · Trial and conversion
  2. How many days into the trial did you see real value?Ruler · Trial and conversion
  3. How confident are you that this is the right tool for your team?Big scale · Trial and conversion
  4. Have you ever needed to find out who changed something?Yes / No · Audit logs and history
  5. How easy was it to find that change in the activity log?Customer effort · Audit logs and history
  6. Which events do you need recorded?Multiple choice · Audit logs and history

Frequently asked questions

Right after the result shows, while the user is still looking at it. From Starter, or on 1 survey on Free, a custom event from your code opens the survey at that moment.

No. Show it until answered, then again after a number of days, so frequent users are asked now and then and the answers still cover many results.

A rising share of results used as they were or edited a little, a rating that holds after model changes, and few low trust scores. Compare each period with the one before rather than with a fixed target.

Yes. Add a second survey as a side button, so users can report a result whenever they spot a problem, and keep the popup for the moment right after a result.

Yes. The same script runs in web apps and single-page apps, and the custom event can be fired from any part of your code.

Yes. Every template comes with every plan, the free plan included, and the free plan covers 150 responses a month across your surveys.

Yes. The template becomes your own survey, so reword any question, add or remove questions, and change the options and the logic. The gallery stays as it is.

Yes, from Starter. The template is written in US English. Add languages by hand or with AI translation, and each person sees the survey in their browser's language.

Listen better, act faster, and build what truly matters.

Start free, see what your customers mean, and get each answer to the person who can act on it.

Get started freeTalk to sales
  • 150 responses a month on Free
  • Every channel and AI analysis
  • No card needed