Typeform recently tested out an AI-powered survey builder that aimed to simplify the form creation process. The idea was to have a chatbot ask the user questions and then auto-generate a survey based on the responses.

My experience

I recently tried out Typeform’s “create with AI” firsthand, answering the chatbot's questions to build a form. 


I understood the survey builder would only give me something to start out with. However, I found the time it took to communicate what you wanted with the AI bot took longer than creating it manually and resulted in a weaker survey. Conclusion: Using the ai form builder simply wasn't worth my time.

Sunsetting their MVP

I assume Typeform found that other users had a similar, lackluster experience as me. Rather than persist with a flawed product, Typeform decided to remove the feature and shift focus elsewhere.

Typeform’s quick test and sunsetting of an unwanted feature provide some excellent lessons on managing innovation.

Too often, companies pour excessive time and money into features nobody asked for. They assume that just because something utilizes exciting new tech like AI, users will use it.

Typeform avoided this trap by first testing a MVP (minimum viable product) before making any grand announcements.

3 Things we can learn from their MVP

  1. The feature was quietly launched to their app to gather real user feedback. It was only discoverable when a user went to build a new survey. This allowed for more limited exposure upon the features release.
  2. They called the feature "beta". This is critical in managing the users’ expectations. If something doesn't work perfectly, the user won’t be displeased because they had lower expectations.
  3. They only kept the feature live for a few months. Typeform wasn't afraid to fail. They pushed out an MVP and let the users validate the feature. Although this experiment wasn't a success, Typeform can try a different experiment instead of spinning its wheels on a “meh” feature.

Thorough validation is required before fully launching any product, no matter how much potential it shows. Typeform not only validated the technical capabilities of AI, but more importantly validated whether the AI creator actually improved the customer experience. 

Your thoughts

What examples of failed experiments have you seen in other products? We are super curious.

(FYI - we ❤️Typeform – check out how we got a free, lifetime subscription)

Your bottom-line called and wants to know how
our Designtific Method can help.

Tell us about your project