Watson Studio
Company
IBM
role
Design Lead
Senior iXD
project type
Platform Initiative
—
Watson Studio exists at the intersection of IBM's data platform and cognitive services. Previously, the two parts of the company existed in silos, and it was part of my team's mission to help define how they should live together. It's purpose is to allow IBM clients and partners to utilize Watson's APIs to build their own domain-specific knowledge models.
a Problem
Regardless of user or use case, all paths start with customers picking one service to solve a specific problem.
However, not all services are unique, and real use cases span multiple technologies. Users are confused at the door. It is hard to pick one service when all of them have similar properties. An API first is a good approach for a developer, but misses the opportunity to sell higher value offerings.
a solution
We propose to start with a new mental model for working with Watson. Instead of service specific, users think of Watson as ONE, and decide which tool to use depending on their data and use case.
By dividing Watson services into three large buckets of "understanding" we can simplify the interaction users have with the AI. They can teach it to understand visual, text-based, and speech information. Within these larger buckets are the different Watson services like visual recognition, natural language understanding, and speech-to-text to name just a few.
High fidelity wireframes based on IBM Design's internal design language Carbon.
In early prototypes, it became very apparent that the tool would become very cumbersome for users without data storage. Thus began the integration of visual recognition with the data platform, eventually being rebranded as Watson Studio. This was the first priority.
The next question we asked ourselves was whether it would be possible to use Watson's base knowledge to help less technical users train Watson. Do you need a data scientist in order to easily aggregate effective data to a model and easily understand how to improve performance?
Once a user has entered Watson Studio's text understanding tool they can begin working with their domain specific data.
Once data has been uploaded, users receive notification that Watson is analyzing their data.
Users are presented with Watson suggestions using a technique that employs the Natural Language Understanding service.
Designs based on future direction of new brand language called Duo.
These future forecasted designs are currently being used to give direction to engineering and offering management to help guide user focus in the roadmap.
More to come…