Build AI Lab feature to demonstrate NetBase’s best-in-class machine and deep learning. It's the fastest and smartest way to discover insights from the public web.
Deliverables
Web feature
Skills and Responsibility
Defining Problem, Market Research, Personas, Competitive Survey, User Flows, Wireframes, Branding, High Fidelity Prototypes, User Testing
Tools and Software
Sketch, Invision, Adobe CC, Principle
the challenge
After talking to users, industry experts, engineers and company executives following information surfaced:
For expert users: Social web contains vast of amount of insights, but the sheer volume of data makes it hard to separate signal from the noise.
For novice users :Tools for social analysis are difficult to use and even harder to master. The better the tool, the tougher it is to use.
Technical issues: Implementing new technology using old frameworks was hard. Additionally at this stage there could only be 70% data accuracy with AI.
Business Perspective: There is a huge market for AI drive social insights. We needed to produce data proof before investment.
User Research
Pain point:
Data analysis is painful.
Pain point:
Tool is useful if you have a problem statement to begin. However it doesn’t surface valuable information about users or crisis etc.
Expectation:
Unstructured data analysis and real-time crisis management.
COMPETITIVE ANALYSIS
I analyzed our competitors to find out the strengths and weaknesses of our competition, so that we could make knowledgeable decisions about their own strategies.
Strengths:
- Well branded
- Well thought usage of AI technology
Weaknesses:
- Expensive
- No structural UI to use AI
Opportunities:
- Be the first in market to launch
- Ability to learn fast and improve
Threats:
- Our AI technology still in learning face
- Little information about what’s most useful
Concept ideation
Lab is a place where ideas are created, tested out and either validated or discarded. Our AI technology is still in its infant stage but we still wanted to showcase it. So the idea of 'lab' was conceptualized.
Our users could see AI in action here, provide us with feedback and improve the technology. Since the users would know this is in the beginning face it would cause less user frustration even if we had only 70% data accuracy.
USER FLOW
With the above information, aim was to create intuitive flows that would help users complete critical routes with utmost ease.
Low and HIGH FIDELITY WIREFRAME
I created these wireframes to ensure that each element on the screen has a purpose, achieves the goals set and define a logical navigation for the feature.
BRANDING ELEMENTS
Logo:
A futuristic looking logo that represents data analysis. Besides being responsive, it should be memorable and contribute to creating an AI Identity.
Typography:
I decided to use Lato, since it is clean, modern, professional yet friendly and open. Additionally, it can be used over 125+ languages.
Colors:
Black being the primary color of our application, I decided to pair it with more futuristic techno neon colors that impart convey precision, accuracy
& dynamism.
And finally......
Once all the basic elements are in place and user tests have confirmed the best possible interaction for the users, it is time to put all the pieces together. Below are some of the screens from the feature.
USER TESTS
Need for help text for users to know what to expect and feedback form to provide input.
Long wait time. Maybe educating users in the meanwhile to the steps AI is taking to complete the task might reduce user frustration.
Contrast of colors make it hard for over 45+ users to read. Either change white to light gray or provide a white theme AI Lab.
