Toggle screenreader descriptions↤
[Image description]: a metallic, bipedal angular figure walks from the right-hand side of the shot towards the left. Its head is hexagonal and its face resembles a single simplified eye.
Alibaba Cloud is the cloud computing arm of Alibaba Group. It provides a comprehensive suite of global cloud computing services to power both international customers’ online businesses and Alibaba Group’s own e-commerce ecosystem.
At the heart of this ambition lies ET Brain, best described as an intelligent, adaptive technology platform for practical applications where data processing, machine learning and real-time processing are key.
This branding task was a little different than others: rather than just creating a graphic identity for a company, we were tasked with turning it into a character.
Alibaba group already had a starting point for us to go from: two brackets with a line between them (symbolising the computing behaviour of "input > computation > output"). The inside of these brackets were stylised into a rounded hexagon of negative space, with the line moved slightly down from the centrepoint to resemble a mouth. This was apparently inspired by the face of their co-founder, Jack Ma.
[image description]: Three images, from left to right.
First image: two rounded square brackets with a thin line between them on a white background. The words "Alibaba Cloud" written below.
Second image: A photo of Jack Ma, the Alibaba Group co-founder with a dotted line outlining the shape of his face.
Third image: The outline of a roudned hexagon with a small circle in the center. The edges of the shapes look soft and faintly glow light blue. They are hovering in place against a dark purple-to-blue background.
Our response was to simplify this into something more concrete: turn the implied face into something that could be read instantly. We replaced the line/mouth with a dot that could be read more like an eye, as we felt that a face with en eye that could emote would be more approachable than a face with just a mouth.
This simple shape acts as a mark that can be identified at large and small scale, as well as used as an element to be incorporated into image compositions.
When we developed the mark, we experimented with filling in the negative space, but opted to just join up the frame and leave the middle empty. This came from a series of motion experiments in which we decided to play with the ethereal nature of an AI and turn the hexagon into various objects that would only be readable as hexagons from specific angles.
[image description]: Three rounds of hand-drawn sketches, interpreting the hexagonal shape in 3D space from different perspectives. The first round shows it framing a box, the second shows the frame on its own and the third shows the hexagon inside of a transparent cube.
Not only did this interpretation of the mark have to exist in virtual space – it also had to work as a physical object. In China, it is common practice for businesses to have small models of their company character to give out at meetings with their most esteemed clients. At this point, the project was split up amongst the team members – bringing the character to life was my responsibility.
[image description]: an assortment of 3D renders, all experimenting with how to turn the ET Brain mark into a physical object.
One of the big questions was whether they wanted it to be more of a mascot, or more of a simple desk object, so a lot of the process involved experimenting with the different ways in which the mark could be extended into the physical world.
Our favourite of these initial routes was the "pebble" (seen on the right-hand side of the collage above), which played into the geometric form of the mark to create objects that would snap together with small magnets. We found this worked particularly well because the client wanted the mascot to be recolourable and reskinnable for different events, partnership and limited editions, so making the objects snap together added to their collectability.
The client liked this, but said that it needed to be more human, suggesting that taking the approach of a more traditional mascot would work best, especially with the brand launch coinciding with the 2018 Winter Olympics, for which they were the main sponsor.
[image description]: eight hand-drawn illustrations of the same polygonal robotic/humanoid form. Each version is covered in a different pattern. Its head is a 3D interpretation of the hexagonal mark, the body and arms are individual floating parts with large, almost rectangular hands and the legs end in rounded points, to suggest that the form floats, rather than stands.
[image description]: A series of hand-drawn sketches showing the development of the ET Character, resulting in the final proportional diagram. The diagram shows a polygonal bipedal figure with a hexagonal head featuring a circle in the centre. The body is the same height as the head, but a lot thinner. The neck is the thinnest part, sloping outwards towards two flat, slightly pointed feet. The figure has no arms.
Finding a body for ET was tricky. It had to be a good balance between human and technological; intelligent and approachable. It had to be simple enough to apply different colours and patterns to, but not so simple that it didn't feel like something special.
[image description]: Six 3D renders of the ET Brain mascot model in different poses. These poses include walking from left to right, looking upwards with its back arched, looking towards the viewer with its head tilted to the right, looking down with bent knees at a spot on the ground, looking slightly upwards to the left and raising itself on one leg, as if it is about to jump.
What worked out the best was a short figure with a large head, emphasising the mark and suggesting intelligence while still appearing cute and friendly. Taking away the arms helped to distance ET Brain from its incarnation in earlier sketches, where its proportions and body language were a little too humanoid.
[image description]: A screenshot of the 3D animation program "Blender." In the window is the ET figure with no lighting or texturing. Through the figure, the viewer can see the "rig" of the model, which is used as an armature to pose and animate the figure.
Since the face was so minimal, body language still played a very important part in conveying emotion. To keep these expressions in line with the design choices that lead us to this form, we made sure to keep ET Brain's physical behaviours upbeat, optimistic and curious.
[image description]: A male Alibaba Cloud employee stands on stage at a computing conference in China. Behind him is a silver ET mascot on a blue background looking up and to the left.
The mascot played a large part in the reveal of the new brand, taking centre stage at The Computing Conference in Beijing and CES in Las Vegas.
[image description]: A render of two ET mascot models on skis, one in gold and one in white with blue/purple gradient detailing.
[image description]: A physical model of the Skiing ET mascot on a table in front of a blue and white box. The box reads: "The 2018 Winter Olympics. Hello Pyeong Chang!"
[image description]: A red metallic physical model in black foam casing inside a red matte box with a gold clip in the shape of the hexagonal mark.
ET Brain became the mascot for Alibaba Cloud's sponsorship of the 2018 Winter Olympics and, as a result, themed models were created. Alongside these, limited edition red metal figures were created for Olympics business partners and competing athletes.
While working on the visuals, I also worked on the sonic identity. Just as the visual brief was a departure from the norm, so was the audio brief. Rather than creating music for videos or logo resolves, my task was to create a voice for the AI.
Sitting on top of an ambient soundscape, different looping layers of audio fade in and out to convey when the AI is thinking, speaking or otherwise processing information.
Because it was multilingual, I strayed away from using anything that resembled speech too clearly, other than vague inflections. For example, the sound of the bouncing dot rises up from the root note of a major scale to the fifth interval, before resting on the third. This mimics the inflection of a human voice asking a question (rising), before resolving to a solution (moving slightly downwards).
The Sound it makes while analysing information from its database features multiple short, echoing arpeggiated notes panning from all directions, suggesting that there is information being brought in at a rapid speed. When it speaks, a filter is used over a more complex melody, creating a sound similar to a mouth opening and closing.
These sounds are played live as a user interacts with the AI and all textual information is displayed on screen as subtitles in the user's chosen language.
These sonic elements were also used in the introduction video, found at the top of this page.