That could have been done in any 3d program, or even after effects(with a little extra work). Blender, Maya, C4D, Max, XSI, etc. They can all make those thought bubble plates easily. Then the video and titles would be composited with something like AE.
AE has now more than enough capabilities to do that kind of simple 3d. Use PSD to create the 3d objects, open the 3d psd layers in AE and animate, or use the more elegant and powerful solution from Zaxwerks- Invigorator ( I recommend the Pro version), where you can import simple vector shapes from AI, and create the 3d elements, then link those elements to a 3d layer in AE and animate at will.