Some Observations

Now that I’ve spent some time with the little guy, and in no particular order:

  • The screen-based UI - When navigating the menu you get kicked out to the eye instead of back to the home screen. This can be frustrating when switching between Home/Settings Menu items.

  • Jibo should remain actively listening after a response so that he might respond to a thank you or whatever. This would certainly make it seem more like a he.

  • When commanded to turn, Jibo always spins counterclockwise. The direction of the spin should be random. He should also be able to spin in either direction on command (left, right, clockwise, counterclockwise).

  • Could be better with natural language overall. Oftentimes I discover that only one specific word works for whatever command. The above is a good example, not just turn “left” but he should also be aware of “counterclockwise.” Or in the menu, not only “scroll to the left,” but saying “again” should repeat the last command. So should “go back,” and “previous screen” or just back or previous…

  • The ability to capture a screen shot might be useful as well.

That’s all I have for now. Overall what Jibo does he seems to do pretty well. He just needs a little polish and to be enhanced by some core skills, specifically for music, (Google) calendar, and traffic reporting abilities. I have no idea what your most requested items are, but these are at the top of my list.



I second Freddy-D’s refinement suggestions.

And Outlook Calendar / Office 365 :wink:

Like particularly to the second point. If Jibo is supposed to be my buddy, he needs to keep the conversation going.

This point is tricky because I think the intent is to give firm control over when Jibo accesses the cloud for speech recognition. Perhaps certain words in addition to his activation phrase could be processed internally (and disclosed appropriately in the training), maybe when followed by his name.

Even this would only take a conversational interface so far. I also wonder if users would accept the idea of the voice connection being turned on and off as needed during a single interaction, if the interaction always started with the activation phrase, the blue ring always came on, and perhaps a specific internally processed phrase could stop the entire interaction, requiring a new activation phrase before speech is sent to the cloud.


There are at least 2 places where Jibo already does this. ( Timer and Alarm ) I hope that this expands to every interaction. Just a couple seconds after the action would be enough.

1 Like

At the very least he should occasionally respond, “you’re welcome,” (or, you’re welcome [insert name]) to a thank you or thanks. and/or respond like, “will that be all?” and then wait for the next verbal command. Those four words would definitely take the experience to the next level, “will that be all?”


I definitely agree that Jibo should actively respond to a “thank you” in most question/answer and command-style interactions in the same way he responds to “Hey Jibo, thank you”. If the old SDK is any indicator of how things are currently setup, it looks like Jibo has this in his blood, so hopefully we can see this in a future update.


I totally agree that this needs to be looked at to make his actions more random. Other suggestions such as calming down, and other commands that while his ring is blue would allow input without having to say his command inputs would be a great help in his growth.


It would be great if he was more interactive with conversations