Visual Verification Feature Update 🔍

Hi everyone, today I'm back with another video to give a quick update on one of our live vision features.

So today we're making visual verification available to all customers.

This feature has been in the platform for, I think, almost half a year now, but only on request for customers when they wanted to test out this feature.

And as of today, we're rolling out this feature to all clients, available in all their new studies.

So I'll just give a quick update on how the feature works, what you can do with it, and what it looks like in the platform.

So as you can see here, I have a study guide open, and I have here an empty question.

And here, when you click on the 3 dots, you have the live vision features where you can add live vision.

And one of the options we have here is the visual verification.

And so visual verification, when you turn it on, it gives you a brief explanation of what you can do, what you cannot do.

For example, you cannot look at people's screen, it's only through the camera.

and also it only looks at visuals, right?

So it cannot look at something that's said out loud, or that people feel, something like that.

so yeah, you can turn on the feature, and then as soon as you turn it on, there is an extra box here at the bottom.

And basically now you can ask people to do something or to show something, you can still give the follow-up probing as you used to, but you can also now ask the AI or the moderator to verify something during the interview.

This allows you to verify something while the interview is happening, and the moderator will also intervene if people do not do it correctly and ask them, for example, to show it again or to film it better so it's better on screen.

and this makes it so that you can verify certain visual things.

without having to to wait until the interview is done and then to see that people might have misunderstood the task or filmed the objects incorrectly or not clear enough.

So you can here you have a quick example.

For example, people were asked, could you grab a mug or a cup or usually drink and hold it up to the camera?

That's what the instructions that the participants get, and then the AI moderator is asked to visually verify on camera if they have a mug or a cup.

and you can write here whatever you want.

It's this is how the default setting is.

You can also end it on a repeated failure.

So for example, here at the bottom we have a a question where they had to show a drinking glass with "end on repeated failure" turned on, which means that if people fail to show the object after 3 attempts- so they get asked to show it, then they fail, so they have to show it again, show it again, and eventually they still didn't show it- then the interview will end.

And so this is something that you can use to to screen out people based on some live visual information.

So that's how you set it up in the study guide.

if you then go to this study, we did here a few example interviews, and we go look at how this looks like on the interview detail page.

You see now we have a few more visual checks.

We already did the live vision quality checks where we check if people are correctly in a in a good scenario and good setting to to start an interview.

So for example here she was not filming herself, so then she's asked to turn the camera around before the interview even starts.

So that's the general quality checks that run on all interviews.

But now we have all our questions here where they were asked to show certain things.

And so for example here at question 6 is a good example.

they were visually verified to show something from the coffee that they drink.

And so in this case, in her first attempt, she did not show it clearly enough.

And then you can open up what did she show.

And for example, in these 3 frames, you can see she films it, but only 1 frame is actually clear clear enough.

And so for the moderator, this is not enough evidence and it's not long enough.

So it will ask again.

And then so in this case, could you hold up the coffee?

And you can see that this is asked again based on the visual verification that happens.

And then in this case, she showed it again.

And then we can see the attempts here where there are more than enough frames.

That show the coffee that she's using.

And so without this visual verification feature, we might have only had one short frame where we could see the coffee, and now we have a clear view of what the coffee was, and we can extract all the information from that.

And so it is just in general make sure that people understand the task correctly, film the things correctly.

Sometimes people forget to flip their camera, things like that can now be prevented on all your interviews.

And so, yeah, that's the visual verification feature.

It's now available for all clients.

you can specify what you want to verify.

people will be asked again if they do not show the objects correctly enough, and you can even screen out people from the interview if they fail to to show the objects or or or the things in camera.

Okay, that's it.

See you in the next video.