How to add tasks to your research study

When you create a research study with Typeform’s Research Flow, you’ll be able to add tasks for your AI-moderated research study to perform. Tasks will help guide the conversation between the AI-moderator and the participant.

You can add the following tasks to your research study:

Welcome screen - welcomes participants to the study
Screener - filters participants based on specific criteria or requirements you set
Greeting - greets participants and provides context to the study
Research questions - collects specific information you need from participants
Ending - wraps up the study with a personalized goodbye

In the article below, you’ll find a breakdown of how each task works and how to add them.

If you’re interested in Typeform’s Research Flow, please reach out to our Sales team.

How to add tasks

From the content panel of your research study, click + Add tasks.

Then select a task you want to add to your research study.

Tasks will run in the order you see in the left side menu.

Welcome screen

A welcome screen is automatically added to your research study by default.

The welcome screen prepares the participant before the AI-moderated conversation begins. It can set expectations, show your branding, and provides a CTA (call to action) button to start the research study. The text on your welcome screen will remain static.

Coming soon: We’re currently working on support for text formatting, including bold, italic, underline and hyperlinks.

In the settings menu on the right, you can edit the text on the button and add an image or video.

When you add an image or video, a window will appear where you can:

Upload your own image (JPG, PNG, or GIF up to 4MB)
Use an image from the Unsplash library
Select an image from My gallery (images you’ve previously used)

Under the Layouts settings, click on the mobile and desktop dropdowns and select which image layout you want for the different devices.

Click the change icon to change the image, or click the trash can icon to remove the image.

Screener

The screener task will ask participants pre-session questions to qualify or disqualify them from the study. The study will end for disqualified participants and the AI-moderator will inform the participant that they are not qualified.

Enter a screening question in the text field and click + Add.

Select the answer format from the dropdown. You can select Multiple choice or Dropdown.

Next, enter your answer choices and select from the dropdowns if the answer choice qualifies or disqualifies the participant. Click + Add choice if you need to add more answer choices.

Click Save when you’re done.

Hover over the question and click the pencil icon to edit the question or click the trash can icon to delete it.

Add additional screening questions by repeating the steps above.

Qualified participants’ responses to screening questions appear in the Results Sessions table. Disqualified participants’ responses to screening questions do not appear in the Results Sessions table and are not included in the Insights panel.

Greeting

The greeting task sets the AI moderator’s opening message to participants and provides context to the research study. The greeting tasks is AI-generated and not static like the Welcome screen.

You can optionally include the following in your greeting tasks:

A brief intro about your company or project
A brief intro about your research
Ask for the participant’s first name
Custom instructions

Click on the checkbox next to the items you want to include in your greeting task.

If you’ve selected:

A brief intro about your company or project
A brief intro about your research

You can click the Edit link to add information about your company, project or research. Or edit any information you’ve previously provided about your company, project, or research when you trained your study.

If you’ve selected to add Custom instructions… you’ll need to enter instructions for the AI moderator to follow. Don’t include questions in the custom instructions as greetings are only for introductions and statements.

Ask research questions

Research questions define the key topics the AI moderator will explore with participants. They help guide each conversation toward the insights you need and make findings more consistent across sessions.

After you provide context about your study, AI will automatically generate a set of research questions. Use these questions as a starting point: review, edit, remove, or add questions to tailor them to your research goals.

To edit a question, hover over the question and click the 3 dots. Then select Question settings to edit it or click Delete to remove the question.

To delete all the research questions, click the trash can icon in the Research questions heading.

To add more questions, type your question in the field and click the + Add button.

A new window will appear with your question, where you can configure its settings.

In the Follow-up questions dropdown, you can choose if you would like the AI-moderator to ask participants follow-up questions based on the answer they give. If a participant provides a vague answer, you can use follow-up questions to clarify their response and gather additional details. You can select to have:

No follow-up questions - the AI moderator asks the research question once and moves on.

Up to 3 follow-up questions - the AI moderator may ask up to 3 AI-generated probing questions based on the participant’s answer.

It’s best to use follow-up questions on open-text questions where you want more in depth responses. The follow-up questions generated will avoid leading questions that nudge participants toward a desired answer. Try to avoid using follow-up questions on structured multiple-choice questions, since the answer options are already predefined.

In the Answer format dropdown, select how you would like participants to answer your question.

Below is a breakdown of the different answer format options.

Answer format	Best used for	Data you get
Open text	When you want qualitative responses to understand the “why” and “how” behind answers. Allows participants to answer in their own words without fixed choices.	Transcript text AI topic/sentiment analysis
Multiple choice	Gathering quantitative data by providing predefined answer options when all possible answers are known.	Bar chart Count Percentage per option
Dropdown	Gathering quantitative data by providing predefined answer options when all possible answers are known.	Bar chart Count Percentage per option

When you select Multiple choice as an answer format, you can also toggle on Multiple selection to allow respondents to choose more than one option.

You can also include a visual stimulus with a research question so participants can react to an image during the session. This is especially useful when you want feedback on something visual, such as packaging, branding, or other design concepts.

Click the + button next to Stimulus image and upload an image.

You can add a stimulus image on any research question but not on screener questions. Each research question can have its own unique stimulus image. When you upload a stimulus image, the AI moderator will analyze the image to generate context that it’ll use during the session when asking follow-up questions. The stimulus image will also be used as context when the session data is analyzed for the insights report.

Click Save when you’re done.

Usability test

The usability test task lets you ask your participants to complete a task on your website. Participants will be asked to record and share their screen while they complete the task and talk through their thoughts. You’ll be able to observe what they’re doing while completing the task and identify usability or discoverability issues in real-time.

Make sure to keep the following in mind while setting up your usability test task:

Use realistic scenarios: Create usability test tasks with real-world user goals and behaviors.
Provide clear instructions: Give direct guidance without telling the participant how to click or navigate on the website.
Make your prompts action-oriented: Use clear verbs to prompt interaction, for example, find, buy, or sign up.
Use unbiased phrasing: Avoid giving away UI clues or answers to your usability test question.

Some examples of a goal for a usability test task could be the following:

“Find and buy a pair of running shoes under $100.”
“Look up the results of your midterm exams.”
“Sign up for a free session at the UX summit.”
“Locate the nearest physical store using the map feature.”

Note that audio responses and screen share recordings are required for this task. Click to find out more about enabling camera and microphone access for your respondents in research studies. The usability test task is only available on desktop devices.

In the Respondent instructions section, you can provide detailed information to help your respondents understand the task they need to complete.

Give your task a title, then explain the steps participants should follow to successfully complete the usability test. Include these instructions before they open the task URL. The instructions provided to your respondents can be up to 700 characters long.

Add a link to your website in the Website URL field. People will be able to click this link directly from your research study to complete the usability test task.

In the Task success for AI moderator section, you can add information that will help the AI determine whether a task was completed correctly by describing what a successful outcome looks like for that task.

Task description: Use this field to explain what the task is and what steps are needed to complete the task.
Final page description: Use this to explain what will appear on the last screen of the task. This will help the AI assistant understand when the task is ending.

Next, you can add post-task questions after the usability test task to better understand how participants felt about completing the usability test.

The first question, asking whether people were able to complete the task, is required and can’t be edited. The answer format of this question is Multiple choice, and the answer options are Yes, No, and Not sure.

You can also select from other answer types to get respondents’ opinions on the task, like Opinion scale, Rating, and Ranking. Click Add to add a new post-task question, or click the three dots to delete it or go to its Question settings.

In Question settings, you can edit your question, allow the AI moderator to ask follow-up questions, and change the answer format. You can also select the scale range for Opinion scale questions, like in the example below, and add labels to different values on the scale. Click Save when you’re done with your selection.

You can click the eye icon next to a question to add display logic to provide visibility for both questions and answers.

This will allow you to choose which questions to show and which to hide based on the answers to a previous question.

For example, you can choose to skip the question asking people what they liked about the process if they gave a low rating, and go straight to the question asking about what they didn’t like about it.

Ending

The Ending task wraps up the study with a personalized goodbye. You can also include a link to a website for participants to visit.

To add a custom farewell message, turn on Custom instructions and enter the message you’d like participants to receive. Keep the farewell message free of questions.

If you leave Custom instructions toggled off, the AI-moderator will generate a farewell message for you.

Toggle on Button if you would like to include a CTA (call to action) button with a link to a website. In the Label field, enter the text you want to appear on the button and in the URL field, enter a link to a website.